Copyright by Amanda C. Skuldt 2013

The Dissertation Committee for Amanda C. Skuldt Certifies that this is the approved version of the following dissertation:

State Sponsored Terrorism? Leader Survival and the Foreign Policy of Fear

Committee:

Peter Trubowitz, Co-Supervisor

Ami Pedahzur, Co-Supervisor

Catherine Boone

Daniel Byman

Daniel Powers State Sponsored Terrorism? Leader Survival and the Foreign Policy of Fear

by

Amanda C. Skuldt, B.A.

Dissertation Presented to the Faculty of the Graduate School of The University of Texas at Austin in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

The University of Texas at Austin May 2013

Dedication

This dissertation is dedicated to my son, Finn Skuldt with great hopes that his path is filled with at least as many great adventures and willing companions as mine has been.

And to the memories of Colonel Lee Charles and Leah Paxton Black

Acknowledgements

The people that have been instrumental in my completion of this dissertation as advisors, critics, cheerleaders, and sounding boards are many, and I am grateful for this opportunity thank many of them by name. This project never would have gotten off the ground if not for the mentorship and support of Ami Pedahzur. I am very grateful for the opportunity I have had to work with him over the years as his teaching assistant and project director in the Terrorists, Insurgents, and Guerillas in Education and Research (TIGER) lab at the University of Texas-Austin. His investment in the success of his graduate students and his drive to tackle the hard questions and do the hard work to understand the dynamics between terrorism and political violence is truly an inspiration. I have been particularly fortunate that I have had the opportunity to work with Peter Trubowitz on this project over the two years. His perspectives on my work and willingness to come on board as my co-chair and walk me though the big obstacles to my theory and cases truly made its completion possible. I am also grateful for the insights, direction, training, and encouragement from Catherine Boone, Daniel Powers, and Daniel Byman. This dissertation has also been made possible in part by the Department of Government and the generous support throughout the years for teaching assistant positions, fellowships, research assistants, and grants for conference attendance. I thank Robert Moser and Gary Freedman for enabling this support and Annette Carlile for her reliable help navigating the bureaucracy.

v There are a number of friends and colleagues who have been instrumental in this process. I am particularly grateful for Dua’ Nakhala and Cassy Dorff who have been alongside me throughout this journey from course work, to qualifying exams, to reading countless drafts of my dissertation and helping me prepare for conferences and other venues where I have presented this project. I am also grateful to Ayesha Ray, who has been my sounding board over the last few years as I have tried to work through various problems and have struggled with my internal obstacles to completion. Furthermore, I would not have even dared pursue an academic career had it not been for the generous support and encouragement of Becky Sarah, Cory Skuldt, Toni Amato, Leanne Doherty Mason, and Kirk James Beattie. I am most thankful for my partner, Colleen Tabolt, and for my son Finn, who both patiently found other things to do while I completed this project. Colleen’s love and support has been constant throughout the elations and depressions that are inevitable in this process. She has been my anchor. Her unshakable faith in me is more than I could have ever asked for and I can’t help but look forward to our next adventure because no matter where we end up, if we are together, it will be home. The countless hours my son has logged on the computer next to me while I’ve worked and the hours we’ve spent in coffee shops reading together has not always been the sort of attention that he’s wanted, but I hope that he will realize as he grows that sometimes just being there is enough. One final acknowledgement goes to our many cats that have kept me laughing this whole time, tolerated my pacing, and listened to me ramble on about logistic analysis and the Iran-Iraq war when everyone else in my life was sick to death of hearing about it. I couldn’t have done it without them.

vi State Sponsored Terrorism? Leader Fortification and the Foreign Policy of Fear

Amanda C. Skuldt, PhD The University of Texas at Austin, 2013

Co-supervisors: Ami Pedahzur and Peter Trubowitz

States that sponsor terrorism pose one of the greatest policy and security challenges of the

21st century. Over the past decade, the United States and coalition allies have invested over a trillion dollars ⁠ in the wars in Afghanistan and Iraq, both waged, in part, to end their support for terrorism. Iran’s support for Hezbollah and Hamas makes negotiations over its nuclear program tremendously difficult and the prospect of an Iranian nuclear umbrella, under which these groups could operate, especially concerning. Likewise,

Qaddafi’s overthrow and the siege on Assad’s regime in Syria have both been justified in the context of their historic support for terrorists, as well as the more recent normative concern for the repression of their people.

This paper moves beyond a simple explanation of state sponsorship as covert war or way to persuade target states to concede policy objectives. Rather, it models state sponsored terrorism as a leader survival strategy that leaders choose when facing simultaneous internal and external threats. By investing a portion of the state’s military power outside the control of the military and into terrorist groups and the security services that arm and train them, the leader is able to signal competency to other elites in vii his coalition and insulate himself from existing threats of coup d’état from the military while avoiding defeat in external conflict. Using a newly constructed dataset on state sponsorship that uses the leader-year (1968-2001) as the unit of analysis (N=5139), this study finds that many existing explanations for state sponsorship do not withstand empirical testing and that the combined level of high external threat and elevated threat of coup d’état are key determinants of a leader’s decision to sponsor terrorist groups.

This work has tremendous implications for US security policy as current practices, such as regime-targeted sanctions, may have the unintended effect of increasing the level of threat that the leader experiences and thus the likelihood of state sponsorship. These insights highlight a major reason why military strikes and economic sanctions are less successful than regime change for ending state sponsorship.

Furthermore, it suggests that carefully reducing the external and internal levels of threat the leader faces may be the most effective method to end state sponsorship of terrorism.

viii Table of Contents

List of Tables ...... xiv

List of Figures ...... xv

Chapter 1: Introduction ...... 1 Argument and Basic Assumptions ...... 6 The Nature of State Sponsorship ...... 7 State Sponsored Terrorism as Warfare ...... 8 State Sponsored Terrorism as Diplomacy ...... 10 State Sponsorship and Leader Survival ...... 12 Conclusion ...... 14

Chapter 2: Literature Review ...... 17 Towards a Working Definition of State Sponsorship ...... 18 The Maximalist Approach ...... 19 The Minimalist Approach ...... 20 Inherent Resources of the Sovereign State ...... 23 Protection ...... 24 Access to Territory ...... 25 Money ...... 26 Arms ………………………………………………………………….26 Intelligence ...... 27 Training ...... 28 Diplomatic resources ...... 29 Arriving at a Definition ...... 30 Literature Review ...... 31 External Threat ...... 34 Weak States? ...... 35 Isolated States with Few Resources ...... 38 Costly Signaling ...... 40

ix Domestic Variables ...... 41 Regime type ...... 42 Legitimacy ...... 44 Regime Cohesion ...... 45 Ideology ...... 47 Natural Alliance ...... 48 Lower Transaction Costs ...... 50 Convergence of the Variables ...... 51

Chapter 3: The Theory of Leadership Survival ...... 56 A Key Missing Variable: The Military ...... 58 Omnibalancing and the Military ...... 61 Regime Type, Size of the Winning Coalition, and Support for Terrorism ...65 Terrorist groups, intelligence agencies, and leadership survival ...... 67 Intelligence ...... 68 Power projection ...... 68 Accountability ...... 69 Sponsorship in Democracies ...... 70 Sponsorship in Autocratic Regimes ...... 73 Threat Environment and State Sponsorship ...... 76 High External & High Internal (HEHI) Threat Environment ...... 78 High External & Low Internal (HELI) threat environment ...... 82 Low External & High Internal (LEHI) Threat Environment ...... 85 Low External & Low Internal (LELI) Threat Environment ...... 87 Hypotheses ...... 89 Conclusion ...... 93

Chapter 4: Empirical Support for the Theory of Leadership Survival ...... 94 Lack of existing empirical foundation ...... 94 Existing Datasets ...... 96 ITERATE (1-4) ...... 97 START and the GTD ...... 98 x

Non State Armed Groups (NAGS), their Targets, and Sponsors ...... 99 Non-State Actors and Alliance Behavior Dataset ...... 100 Challenges with Data Collection ...... 102 Case selection and the Unit of Analysis ...... 103 The Dependent Variable: State Sponsorship of Terrorism ...... 105 Qualifying Groups ...... 106 Qualifying Support...... 108 Types of Support ...... 110 Coding the Dependent Variable ...... 113 The Exclusion of “Passive” Support and Private Financing ...... 117 Independent Variables ...... 119 Regime Type ...... 119 Threat Environment ...... 121 Military Expenditure as % of GDP ...... 124 Control Variables ...... 125 Methods of Analysis and Regression Tables ...... 127 Model 1 (HELI only) ...... 128 Model 2 (HEHI only) ...... 131 Model 3 (LEHI Only) ...... 132 Model 4 (LELI Only) ...... 133 Model 5 (HELI & HEHI Only) ...... 134 Model 6 (HELI, HEHI, & LEHI) & Model 7 (HELI, HEHI, & LEHI controlled for superpowers)...... 137 Relevance and Predictive Value ...... 138 Conclusion ...... 140

Chapter 5: Conforming Cases ...... 142 Ayatollah Khomeini’s Iran ...... 145 High External Threat Environment for Iran ...... 146 Nature of the High Internal Threat in Iran ...... 151 Terrorist groups and Leadership survival in Iran ...... 153

xi Saddam Hussein’s Iraq ...... 164 Threat Environment for Saddam Hussein ...... 166 Iran-Iraq War and External Threat to Saddam’s Regime ...... 167 Internal threat and the Iran-Iraq War ...... 169 Terrorist Group support and Regime Survival during the Iran-Iraq War ...... 173 Operation Desert Shield/Desert Storm and Saddam’s External Threat ...... 177 Internal Threat and the Gulf War ...... 179 Terrorist group support and leadership survival in the Gulf War ...... 180 Concluding Remarks about Iraq ...... 186 Mummar Qaddafi’s Libya ...... 187 External Threat Environment in Qaddafi’s Libya ...... 189 Internal Threat Environment and Qaddafi ...... 193 Terrorist groups and Qaddafi’s leadership survival ...... 196 Conclusion ...... 201

Chapter 6: Outlier Cases ...... 203 Case Selection ...... 203 LELI cases of sponsorship ...... 207 Malaysia’s Tun Abdul Razak bin Hussein ...... 207 External Threat Environment ...... 209 Internal Threat Environment ...... 212 Malaysian Support for the MNLF ...... 213 What explains Malaysian support for the MNLF? ...... 214 What about the other LELI cases? ...... 216 HEHI Non-Sponsoring States ...... 219 Alberto Fujimori’s Peru ...... 219 External threat ...... 221 Internal threat ...... 223 Shining Path and MART: Insurgency in the Highlands ...... 224 xii Why didn’t Fujimori support terrorism? ...... 228 Montesinos ...... 229 Conclusion ...... 230

Chapter 7: Implications and Future Research ...... 232 Policy Implications ...... 234 Ending state support for terrorism ...... 236 Pathways for future research ...... 239 Case specific pathways ...... 239 HELI Threat Environments and Sponsorship ...... 241 Effect of State Sponsorship on International Relations ...... 242 Conclusion ...... 243

Appendix A ...... 229

Bibliography ...... 250

xiii List of Tables

Table 2.1: Types of State-Terrorist Relations ...... 21 Table 3.1: Theory predictions for sponsorship in each threat environment ...... 78 Table 4.1: Excerpt from the Leader-year dataset to illustrate coding protocol. ..104 Table 4.2: Threat Environments that leaders face and the occurrence of state sponsorship...... 123 Table 4.3: Regression Models 1-4 fit with each threat environment examined separately...... 130 Table 4.4: Leader Sponsorship and Threat Environment by Regime Type ...... 133 Table 4.5: Fully fit threat environment Models 5, 6 and 7 (that accounts for the Superpowers)...... 136 Table 4.6: Threat Environment for Leaders who’s Countries have been on US State Sponsored Terrorism List and Theory Prediction ...... 139 Table 4.7: Threat Environments and a Sample of Sponsoring Leaders ...... 140

xiv List of Figures

Figure 1.1: Declining sponsorship trends in the Middle East and North Africa ...... 3 Figure 1.2: Slightly inclined trend of state sponsorship in sponsorship in Asia ...... 3 Figure 1.3: Inclined trend in state sponsorship in Sub-Saharan Africa ...... 4 Figure 3.1: Number of Leaders and Sponsors in Democratic Regimes (1968-2001)...... 73 Figure 3.2: Percent of Autocratic Leaders that supported terrorism (1968-2001). 75 Figure 3.3: A Breakdown of the Threat Environment Leaders Faced (1968-2001)76 Figure 3.4: Sponsoring leaders as a total percentage of threat environment facing leaders (1968-2001)...... 80 Figure 3.5: Percent of Democratic Leaders that are State Sponsors, broken down by Threat Environment (1969-2001)...... 82 Figure 3.6: Sponsoring Leaders Facing the HELI Threat Environment broken down by Regime Type (1968-2001), regime type averaged over tenure. ..84 Figure 3.7: Leaders and sponsors in the LEHI threat environment (1968-2001) ..87 Figure 3.8: LELI facing leaders and percent that are sponsors (1968-2001) ...... 88 Figure 3.9 Leaders facing Internal War and their relationship with State Sponsorship 1968-2001...... 91 Figure 4.1: Frequency of the types of support included by scholars of state sponsorship in 23 surveyed works...... 109 Figure 4.2: Binary Categories of Support occurrence in the country-year dataset112 Figure 4.3: Binary State Sponsorship Variable breakdown in the data ...... 116 Figure 5.1: Iranian Regular Army and IRGC force strengths under Khomeini; Source: IISS ...... 150

xv Figure 5.2: Number of terrorist groups supported and coup attempts faced by Khomeini, by year (1979-1989) ...... 155 Figure 5.3: Active terrorist groups supported by Khomeini (1979-1989) ...... 156 Figure 5.4: Hezbollah Attacks (1982-1988) by Target State...... 160 Figure 5.5: Iraqi Regular Army Personnel (1979-1989) ...... 168 Figure 5.6: Coups in Iraq prior to Saddam assuming the Presidency ...... 170 Figure 5.7: Strength of the Republican Guard (in brigades) and Coup attempts against Saddam Hussein (1979-1989) ...... 172 Figure 5.8: Number of attacks perpetrated by select groups Saddam was supporting during the Iran-Iraq war...... 174 Figure 5.9: Support given the MeK by Saddam Hussein during the Iran-Iraq war (1979-1989) ...... 175 Figure 5.10: Target patterns by the MeK inside Iran during the Iran-Iraq war; ..176 Figure 5.11: Number of Iraqi Regular Army Personnel (1979-2003)...... 179 Figure 5.12: Attacks perpetrated in Iran by the MeK (1990-2002) ...... 181 Figure 5.13: PKK targets and number of attacks while receiving support from Saddam Hussein (1990-1995) ...... 184 Figure 5.14: Attacks by Abu Nidal and the PFLP while supported by Saddam Hussein (1990-1997) ...... 185 Figure 5.15: The ratio of air force personnel to combat aircraft in Libya (1971-1990); Source: IISS 1971-1990 ...... 192 Figure 5.16: The relationship between threat of coup (plots and attempts), MIDs, and number of sponsored active terrorist groups in Qaddafi’s Libya. ..199 Figure 6.1: Military Expenditure as % of GDP under Abdul Razak ...... 210

xvi Figure 6.2: Terrorist attacks by the Shining Path and MART in Fujimori’s Peru (1990-1999)...... 226

xvii Chapter 1: Introduction States that sponsor terrorism pose one of the greatest policy and security challenges of the 21st century. As early as the late 1960s, state sponsored terrorist groups began hijacking and planting bombs aboard aircraft and attacking airports and embassies. These attacks instigated countless diplomatic crises between states, were responsible for massive governmental restructuring, the creation of special forces and counter-terrorism units, prompted reprisal attacks, invasions, and occupations, and have added a level of uncertainty and fear to events such as taking an airplane trip to visit family or staying a night in a hotel.

Several of the US government’s current top foreign policy challenges are directly related to state sponsored terrorism. For example, Iran’s relationship with Hezbollah and

Hamas adds a complicating dimension to the negotiations over its nuclear program. In addition, the United States and it’s coalition allies have invested nearly US$1.5 trillion dollars1 in the eradication of the Taliban regime in Afghanistan that harbored al-Qaeda and on the regime change in Iraq--legitimized in part by the fear of Saddam transferring weapons of mass destruction to terrorists (Bush, State of the Union 2002). Furthermore, states that are the targets of terrorist attacks are more likely to enact domestic political changes that restrict the freedom of the population and the media, pushing policy makers into more aggressive postures and increasing international tensions (Donohue 2008; Pape

2003; Weinberg 2008; Wilkinson 1977, 2006).

1 The US Congressional Research Service estimates that by 2011, US$1.23 trillion has been spent by the US alone. See Amy Belasco, “The Cost of Iraq, Afghanistan, and other Global War on Terror Operation 1 All of the major terrorist attacks against the United States and its assets abroad, including the attacks of September 11th, have been state sponsored. Worldwide, state sponsored groups have been responsible for more than 23,000 attacks resulting in over

63,000 fatalities.2 Furthermore, there is a mounting body of evidence indicating that terrorist groups who receive state sponsorship launch attacks with far greater lethality and frequency than those not backed by the resources of a state (Byman 2005; Enders &

Sandler 2002; Hoffman 2006; see also Cronin 2002; and Quillen 2002). Indeed, despite the appearance of independent financiers of terrorism, such as Osama bin Laden, or remittance and charity funded networks, states still play a major role in the perpetration of all terrorist attacks worldwide.

The clandestine nature of state sponsorship has been a serious obstacle for the advancement of scholarship. Despite its prominence as a highly politicized foreign policy concern during the Cold War and its persistence today, we know very little about it. As the casualty count continues to climb from both state sponsored terrorist attacks and retaliatory military strikes, important questions remain as to how to best understand state sponsorship of terrorism and how to most effectively counter it.

Importantly, while state sponsorship in regions like the Middle East and North

Africa is in decline (see Figure 1.1), the trend in Asia (Figure 1.2) and Sub-Saharan

Africa (Figure 1.3) show a different picture of the phenomenon. In both Asia and Sub-

Saharan Africa, sponsorship is holding steady and even appears to be increasing:

2 Data covers 1968-2008 and was compiled by the author from the GTD and the original dataset for this project. Listed totals include military and non-military targets. The number of attacks for non-military targets only is 18,440 and the number of non-military fatalities is 40,696. 2

Figure 1.1: Declining sponsorship trends in the Middle East and North Africa

Figure 1.2: Slightly inclined trend of state sponsorship in sponsorship in Asia

3

Figure 1.3: Inclined trend in state sponsorship in Sub-Saharan Africa

These trends bring into sharp focus the need to make real headway in understanding the causes of state sponsorship, especially in regions of the world where sponsorship is understudied or even ignored completely.

Leaders who make the decision to provide support to terrorist groups expose the state to the dangers of retaliation, sanctions, and normative censure (Collins 2004).

Providing support to terrorist groups thus undermines two defining features of the modern nation state: territorial control3 and the monopoly of the legitimate use of force,4

(Thompson 1994; Weber 1919) both of which can threaten the leader’s position of power

3 Territorial control is ceded, in many instances, over areas terrorist groups are given for training camps, bases, etc. In extreme cases, it can result in the creation of a “state within a state” such as Southern Lebanon under the control of the PLO in the 1970s. 4 This study focuses specifically on the monopoly on the legitimate use of external force. 4 and in some cases, very survival. In addition, as levels of support increase, leaders often have a difficult time retaining control over the group, often rendering this a high-risk venture (Bapat 2011; Byman & Kreps 2010). Thus, the fact that some leaders of states make the choice to sponsor terrorism, while others do not, leads to the major question this study addresses: under what conditions will leaders provide terrorist groups with access to the resources of the state, or in other words, sponsor terrorism?

By using the leader, rather than the state, as the unit of analysis, this project is able to move inside the state to look at the various domestic security threats faced by a ruler: from the military, elites, and the population, in conjunction with the external threats they are acting against. A central argument of the dissertation is that the leader has to appease or balance against multiple sources and levels of threat and that the terrorist groups provides the leader a unique opportunity to respond to, or circumvent some of these threats in order to retain power.

This study moves the trend away from normative and anecdotal accounts of state sponsorship, stepping outside the obsessive focus on relative state strength and regime type and instead argues that a leader’s choice to sponsor terrorist groups spans regime type5 and state size, and seeks to contribute to the growing body of empirical work that objectively examines the incentives states have to sponsor in the context of international relations.

5 Although the theory notes that the incentives and constraints tend to be stronger for autocratic than democratic regimes because of the different roles the military tends to play. 5 ARGUMENT AND BASIC ASSUMPTIONS This project combines the literatures on terrorism, foreign policy, and civil- military relations to advance a new theory of the determinants of state sponsorship of terrorist groups. We begin with four assumptions about the state and domestic politics.

The first assumption is that a leader’s primary objective is to retain power (David 1991;

Nincic 2005; Trubowitz 2010). This preference to secure a hold on power is true for leaders regardless of regime type, but in certain circumstances it can come even at the expense of the state itself (David 1991). Second, this study assumes that the decision to sponsor a terrorist group is strategic and rational (Byman 2005; Hoffman 2006). This implies that rather than situating the study to look for case-specific or personality-based explanations, there is an observable process that combines incentives and constraints to produce a policy choice of state sponsorship of terrorism.

The third assumption that drives this study is that the state is not a unitary actor.

Rather, there are multiple factions within the state that can challenge the leader’s legitimacy and ability to rule, namely the military, other elites, and the population. The assumption that the state is not unitary is widespread in the state sponsored terrorism literature, yet the implications have not been explored adequately. Especially understudied is the role that the military and the intelligence services play in the state sponsor-terrorist group relationship. Lastly, this study incorporates as a final assumption, the insight that all new leaders come to power in the shadow of the failures of previous regimes. In democratic regimes, these failures often were what prompted the incumbent’s loss and the new leader’s electoral victory. While for other leaders, especially those who

6 have seized power via a coup d’état, the policy failure of the previous regime with critical security and/or economic issues often was the cause of the regime collapse or seizure of the state. Thus, regardless of regime type, the shadow of policy failure creates the broad contours of what is required for a new regime’s legitimacy and stability as the leader works to distance his or herself from these policy failures and create alternative means to address critical issues.

These four assumptions are building blocks we use as we move away from envisioning state sponsorship as simply a form of covert warfare or coercive diplomacy and towards a new approach that places regime survival at center stage.

THE NATURE OF STATE SPONSORSHIP Many scholars have conceptualized state sponsored terrorism as being a “type” of political behavior commonly examined under the rubric of foreign policy, specifically warfare (Cline & Alexander 1986; Jenkins 1974; Merari 1993) or diplomacy (Gal-Or

1993; Selth 1988). While other scholars consider state sponsorship to be an alternative for the state to the regular pursuit of war and negotiation (Byman 2005; Hoffman 2006).

Some, however, have classified state sponsorship as a “tragedy of the commons,”

(Conybeare & Sandler 1993) or merely as the “spillover” from a neighboring domestic conflict (Addison & Murshed 2005). Yet, there is general consensus state sponsorship is a kind of foreign policy action (Gal-Or 1993). This study adopts this consensus as a starting point.

This next section examines the debate regarding whether state sponsorship is a type of warfare, a type of diplomacy, or is an alternative to them both. It outlines the 7 predominant understandings of state sponsorship and exposes flaws in these conceptualizations. The last section will demonstrate this debate lacks utility for understanding the basic nature of state sponsorship and will construct an argument for utilizing a regime survival theory to best understand the choices of state leaders to allocate state resources to terrorists.

STATE SPONSORED TERRORISM AS WARFARE The conventional explanation of state sponsorship is that it is a form of, often undeclared, warfare. This explanation is rooted in the theories of low intensity conflict

(LIC) and covert warfare (Cline & Alexander 1986; Prunckun 1997; Smith 2005;

Thompson 1989). The logic of this is simple: the danger that overt aggression and conventional warfare could escalate to a nuclear confrontation in the years following the

Second World War constrained how states dealt with conflicts. A state’s reliance on its conventional forces became significantly less effective for achieving foreign policy objectives than before (Gilpin 1981; Osgood & Tucker 1967). Some scholars, however, view state sponsorship as adaptation of warfare by states, not just in response to the introduction of nuclear weapons, but also to the proliferation and highly destructive power of modern conventional arms (Kegley, Sturgeon & Wittkopf 1984). Because of these changes, states increasingly resorted to the use of proxies in order to meet their foreign policy objectives (for early indications of this trend see Gaucher 1965 and

Jenkins 1974). Indeed, the use of proxy groups in warfare has become widespread. In some cases, these proxies are, and have historically been, terrorist groups (Hoffman

2006). 8 Long before state sponsorship became commonplace in the international system in the 1980s, Brian Jenkins warned that governments might “recognize the achievements of terrorist groups and begin to employ them or their tactics as a means of surrogate warfare against another nation” (1974, 13), becoming a “deliberate instrument of foreign policy: a cost effective means of waging war covertly” (Hoffman 2006, 258). A theory of warfare applied to state sponsored terrorism, however, is not without its problems.

There are two major analytical flaws in this understanding of state sponsorship.

First, state sponsorship of terrorism is conceptualized as a type of war, while simultaneously it is explained as a way to lower the costs and risks of war (Gal-Or 1993).

Thus, it is war that is, via terrorism, not war. Some authors try to avoid this dilemma by arguing that it actually falls short of the threshold of war (see Angstrom 2005) or that rather than warfare, per se, it is a form of “costly signaling” that conveys a message to the adversary about the level of commitment to a state’s foreign policy objectives (Bapat

2011; see also Fearon 1994).6 This conceptualization overestimates the unitary nature of the state and inadequately accounts for the frequency with which state sponsors and target states do wage conventional warfare with each other, often while simultaneously sponsoring terrorism. While effectively capturing the coercive dynamic inherent in state sponsorship, this is-but-isn’t-quite explanation prevents scholars from developing a clear conceptualization of state sponsorship as an international political phenomenon.

6 Interestingly, Bapat (2011) did not explore the possibility that rather than leaders using state sponsorship to signal commitment to foreign policy objectives to their adversaries, they could be signaling their domestic support base instead. 9 Second, a framework that situates state sponsorship solely as a tactic of warfare cannot adequately deal with its use of, and the effect on, international diplomacy, a competing conceptualization. This study argues that while state sponsorship can be, and is, certainly used in the environment of warfare, understanding sponsorship as international warfare risks conflating the context of state sponsorship with its essence and overlooks the other ways in which the state sponsorship of terrorist groups influences the policies of other states and affects international relations in ways that are not related to war. This issue will be addressed in the discussion of coercive diplomacy below.

STATE SPONSORED TERRORISM AS DIPLOMACY A competing way to conceptualize state sponsored terrorism is not as a form of warfare, but rather as a diplomatic tool. Coercive diplomacy is generally understood as a strategy that combines diplomatic engagement with the threat of force as a way to pressure an adversary to stop an action already underway or to take an action desired by the coercing state (George 1992). Some scholars have argued that coercive diplomacy is a much more accurate lens through which to view state sponsored terrorism (Gal-Or 1993).

State sponsored terrorist attacks allow the sponsor to have extraordinary influence on the policies and behavior of the victim and observer states (Livingstone & Arnold 1986), provides an opportunity to enter the diplomatic arena (Gal-Or 1993), and often utilizes the infrastructure of diplomacy, targets diplomatic assets, and threatens diplomacy as an international institution (Selth 1988).

Stohl & Lopez, on the other hand, see state sponsored terrorism not as a form of coercive diplomacy, but rather, see coercive diplomacy as an international manifestation 10 of state sponsored terrorism in that it is an overt attempt “to make noncompliance with a particular demand…‘terrible beyond endurance’” (1988, 4; see also Schelling 1966).

That is, state sponsored violence, or threat of violence, for coercive means is the norm in the international system, not the exception. It is only in particular cases that this violence has been deemed illegitimate.

States that sponsor terrorism are examples of this “illegitimate” violence and tend to be isolated in the international community (Byman 2005). Some scholars argue that as a means to end this isolation states will sponsor terrorism “as a bargaining technique…

[that] may be used not only to achieve a specific goal but to stimulate diplomacy itself, a means of diplomacy aimed at generating further diplomacy” (Gal-Or 1993, 14). On occasion, ‘rogue’ or ‘outlaw’ states have been known to use the crisis of a terrorist attack to catapult themselves into diplomatic interactions with states with which ties were otherwise severed (14). For example, on more than one occasion, Syria is believed to have orchestrated, or at least been intimately involved in, terrorist attacks, specifically those that involved taking hostages, that it then immediately condemned and stepped in to help resolve (Gal-Or 1993).

The framework that views state sponsorship solely as a pathway to diplomatic interactions, however, cannot explain the behavior of states that sponsor terrorist groups, but do so covertly. Indeed, “in almost all incidents of state-sponsored terrorism a major common denominator has been the tendency of the sponsors to avoid responsibility”

(Gal-Or 1993, 12; see also Byman 2005). Nor can it illuminate why states support groups who perpetrate attacks that appear to be contrary to the state’s diplomatic interest (Byman 11 2005), a paradox that has yet to be fully explored in the literature. Thus, while state sponsored terrorism is somehow coercive in nature, and clearly associated with the ability to change the policies of other states it is difficult to see how state sponsored terrorism fits cleanly into the paradigm of coercive diplomacy.

Thus, neither a strict framework of warfare nor one of diplomacy is adequate to explain the state sponsorship of terrorism. Furthermore, both of these models tend to ignore the domestic political environment completely.

STATE SPONSORSHIP AND LEADER SURVIVAL What the theories of state-sponsorship-as-war and state-sponsorship-as-coercive- diplomacy have in common is that they each attempt to frame state sponsorship as a political behavior that works outside normal channels to extract concessions from another state, as both war and diplomacy are common arenas for the state application of coercion

(Schelling 1966; Stohl & Lopez 1988). Although scholars remain conflicted as to exactly what role state sponsorship plays in this coercive capacity, it is “…generally agreed that state sponsored terrorism is an instrument of foreign policy” (Gal-Or 1993, 9). The function of state sponsorship as an instrument of foreign policy is oddly both an unexplored, but fundamental, assumption in the state sponsorship literature and completely overlooked in the literature on foreign policy.

Studies of foreign policy tend to focus on one of three behaviors of states: state expansion (including also over or under expansion), the development of grand strategy, and how states can be expected to react when faced with an external threat (fight, bandwagon, balance, appease, etc.). In all of these situations, leaders are faced with a 12 question of how to use limited means to achieve their political ends. The reality of limited resources requires a trade off between spending on security (guns) and spending on domestic infrastructure and investment in social programs (butter). The decision of where to invest these limited resources is largely the result of domestic political fights and produces winners and losers within the regime, the supporting elites, and the population at large. Ultimately, state sponsorship is best conceptualized within the framework of these trade-offs, threats, and resources specifically how they affect the probability for regime survival.

Existing explanations of state sponsorship tend to focus on strategic and international variables, especially those developed by scholars of international relations.

This dissertation argues that it is also necessary to incorporate issues of legitimacy and other domestic political pressures into the analysis. As this study will demonstrate, it is the convergence of these two realms and the cross-pressures they put on the ability of the regime to retain power that pushes the leader toward the rational decision to sponsor terrorist groups.

To reiterate, state sponsorship is misunderstood when conceptualized only as a form of warfare or coercive diplomacy. This dissertation argues that while state sponsored terrorist groups can be used by states in the context of warfare or diplomacy, the policy choice to provide terrorist groups with state assets is primarily a function of the leader’s attempts to fortify their position and maintain a hold on power in the face of converging domestic and external security threats to the regime.

13 When our four assumptions about the state and domestic politics-- that leaders prioritize regime survival, state sponsorship is a rational and strategic decision, the state is not a unitary actor but rather it contains multiple potential rivals to the leadership, and that leaders live in the shadow of the failure of the old regime—are introduced to the reality of a world of limited resources, multiple security threats, and hard trade-offs between economic goods and security, then a framework begins to emerge that can better explain the complexities and puzzles of state sponsorship.

CONCLUSION This study seeks to make several contributions to the literatures on terrorism, state sponsorship, and foreign policy. First, it will contribute a clear conceptualization of what state sponsored terrorism is and what it is not. State responsibility for the actions of terrorist groups, and the counter-terrorism options available to target states, hinge on the question of exactly what actions constitute state sponsorship. Second, this study will contribute to a growing body of literature that deals with states outsourcing coercive violence to nonstate actors, including private security companies, insurgent groups, and terrorists. Thirdly, it presents a new argument about how the interaction of domestic and international incentives and constraints shapes the menu of foreign policy options for leaders and makes state sponsored terrorism an attractive choice for some of them.

The rest of the study is organized as follows. Chapter 2 defines state sponsored terrorism and proposes a typology of forms. It examines the multitude of existing explanations within the terrorism literature, and an emerging body of work in IR, as to why states sponsor terrorism. It will then argue that the dominant explanations for the 14 causes of state sponsorship cannot stand-alone; rather they must be complemented by a theory of state sponsorship as a strategy of regime survival. The conjunction of strategic and domestic political variables goes far in accounting for the factors that push a new leader to choose of state sponsorship of terrorist groups.

Chapter 3 lays out the rationale for the Theory of Leadership Survival and makes some predictions regarding the conditions under which leaders can be expected to extend their support to terrorist groups. This chapter also derives hypotheses to be tested.

Chapter 4 provides statistical support for the Theory of Leadership Survival. It will show why existing datasets are inadequate for answering questions related to state sponsorship. It will then outline the quantitative portion of the study and present the results of the statistical analysis. It will describe the data collection process for a unique dataset constructed for this project and address the criteria for case selection, methodological trade-offs, measurement of the dependent and independent variables, and other issues of operationalization. It will then explain the methods used to analyze the data, test the hypotheses using the newly constructed dataset on sponsorship by leaders

(N=5132), and present the results of the descriptive and inferential analysis.

Chapter 5 presents some case studies to probe the causal mechanisms and evaluates the hypotheses and the overall theory in relation to the historical record. It explores in depth the decisions to provide sponsorship to terrorist groups by Khomeini’s

Iran, Saddam Hussein’s Iraq, and Qaddafi’s Libya, and frames their choices to support terrorists in the context of the converging domestic and international threats outlined in the Theory of Leadership Survival that was proposed in Chapter 3. 15 Chapter 6 tackles the outliers that potentially challenge the theory. Instances of sponsorship in cases where the theory predicts no sponsorship (Tun bin al Razak’s

Malaysia) and no sponsorship where the theory would predict it (Fujimori’s Peru) are explored. This examination of outliers further serves to explore the causal mechanisms and dynamics at work in the state sponsor-terrorist group relationship.

Finally, Chapter 7 revisits the argument, presents the policy implications that can be drawn from the findings of this study, examines other possible challenges to the argument, and discusses paths for further research.

16 Chapter 2: Literature Review

As it stands in the literature today, the phenomenon we call “state sponsored terrorism” encompasses a very wide range of state behavior in relation to terrorist groups both in their domestic arenas and internationally. Scholars and policy makers alike have attempted to isolate and define the various ways in which states provide multiple types of assistance to terrorist groups, while international law has created an increasingly less stringent series of “tests” for state responsibility (Jinks 2003).7 In more recent years, the concept of state sponsored terrorism has been stretched to include cases where the state has no intention to support terrorist groups, but merely lacks the interest or capacity to prevent their activities (Byman 2008, Piazza 2008). Clearly, then, the first task must be to determine what exactly “state sponsored terrorism” is.

Paul Richards claims, “to understand war we must first deny it special status… but try instead to grasp its character as but one among many different phases or aspects of social reality” (2005, 3). Likewise, in order to understand state sponsored terrorism, we must first deny it is a phenomenon that is somehow “unnatural” in the realm of international political violence. State sponsored terrorism is a broad category within which a wide variety of state behaviors have been lumped. Despite various attempts to enumerate and describe different types of state sponsoring behavior, the term ‘state sponsored terrorism’ continues to suffer from the same definitional and conceptual

7 The ICC case Nicaragua v. The United States codified that a level of “effective control” must be established over terrorist groups in order to be held responsible. This ruling essentially absolved the United States of any responsibility for contra activities because it could not be determined that they were acting as de facto agents of the United States. Following the September 11, 2001 attacks, however, moved the threshold of responsibility from “effective control” to that of merely “harboring” terrorists. 17 problems that plague terrorism studies in general (see Schmid & Jongman 1988;

Weinberg & Pedahzur 2003).

In the West, ‘state sponsor of terrorism’ is a normative and derogatory label used often by policy makers to describe actions by “rogue” states that would be classified differently if these same behaviors were instead carried out by states directly, or if the states providing support were allied with the United States. This discussion illustrates how important it is to begin with two elementary tasks: 1) develop a working definition of state sponsored terrorism; and 2) explain where it falls among the range of policy options available to state leaders. The definition developed here is that state sponsorship is the intentional allocation of particular state controlled resources to a nonstate group that perpetrates terrorist attacks in a foreign country.

TOWARDS A WORKING DEFINITION OF STATE SPONSORSHIP State sponsored terrorism is a term that appears to have static and clear meaning, but has actually evolved over the past 50 years. It is now an umbrella term that includes a large degree of variation in state behavior, or even inaction. Various government agencies, policy makers, and scholars have developed different definitions, making it difficult to clearly assess whether a particular act qualifies as state sponsorship, since depending on the source and their objective, the criteria vary. For example, the US State

Department, responsible for designating and sanctioning state sponsors, considers a state to be a sponsor of terrorism when the Secretary of State determines that the country has

“repeatedly provided support for acts of international terrorism.” (USDS, “State Sponsors of Terrorism”). This definition is especially concerned with the provision of “critical 18 support”--without providing an explanation of what constitutes such support--declaring that without this support, “terrorist groups would have greater difficulty obtaining the funds weapons, materials, and secure areas they require to plan and conduct operations”

(USDS, “Country Reports on Terrorism" 2007). Other agencies, such as the U.S.

Department of Defense, focus more on the definitions of international and domestic terrorism and consider state sponsorship to be simply any support provided to terrorists by sovereign states (Decker 2003). While the above policy definitions are clearly vague, scholars have tried to unpack the term and have offered a number of definitions that tend to fall into one of two categories: maximalist and minimalist. These categories reflect a longstanding debate regarding whether official state participation in terrorist attacks should be included as part of state sponsorship, the maximalist position, or whether the term state sponsorship is specifically reserved to define the action of to a state providing support to nonstate group, which is the approach of the minimalist definitions (Gal-Or

1993).

THE MAXIMALIST APPROACH Maximalist definitions include the full spectrum of state involvement in terror- related political violence: internal repression, support for external non-state actors, and the use of state officials to target dissidents or other political targets outside the state

(Stohl & Lopez 1988). In part, the maximalist definition is left over from the time before scholars had differentiated repressive state terrorism from the state sponsorship of terrorist groups. Cline & Alexander (1986) outline a common maximalist definition of state sponsored terrorism as: 19 (T)he direct or indirect instigation by a government of official and non-official groups to exercise psychological or physical violence against political opponents, another government, or other entity for purposes of coercion and widespread intimidation to bring about a desired political or strategic objective (46).

Maximalist definitions of state sponsorship includes terrorist state behaviors that are outside the scope of this study, such as attacks by Libyan government officials on dissidents abroad, the use of subversion or support of coups by the CIA or the KGB, and the targeted killings of Palestinian militants by the Israeli Mossad. Maximalist definitions are in danger of over-incorporating acts of political violence and repressive regime terror that occurs within states into their understanding of state sponsorship—including instances of ethnic cleansing, genocide, or widespread targeting of political opposition and this approach complicates the discussion rather than help to clarify it.

THE MINIMALIST APPROACH Minimalist definitions, in contrast, exclude domestic terrorist challengers to the regime except in circumstances where they are backed by an outside state (Byman 2005;

Cline & Alexander 1986; Stohl & Lopez 1988). Many of these definitions also incorporate the idea of the intentional provision of resources such as “a government’s intentional assistance to a terrorist group to help it use violence, bolster its political activities, or sustain the organization” (Byman 2005, 10). Specifically, the minimalist definition is concerned with delineating the relationship between the state and the nonstate actor (Gal-Or 1993). This study uses the assumptions of the minimalist approach to narrow the parameters and assist in developing a clear working definition of state sponsorship.

20 This position not only helps to restrict the relationship to one between the state and the terrorist group, but also keeps the focus on intentional (active) assistance, rather than stretching it to include passive support. However, this definition does not provide information about, or put bounds on, the types of “assistance” offered, only on its function, nor does it allow us to differentiate between assistance to domestic and foreign terrorist groups—a task that will be undertaken below.

It is vital that state sponsorship of terrorism be differentiated from state terror and other proximate behaviors of the state. There are two fundamental characteristics that state sponsorship of terrorism has that are not shared by other forms of coercive, even terrorist, violence by the state. First, state sponsored terrorism is a foreign policy (Gal-Or

1993), and second, state sponsored terrorism is a relationship specifically between the state and a nonstate terrorist group. The variety of possible “state-terrorist” relations are illustrated by Table 2.1 below, which shows that only one type of this relationship is state sponsored terrorism. Importantly, in the following scenarios, it is assumed that State A sponsors Terrorist Group b:

Attack perpetrated by: Attack occurs: Agents of State A Terrorist Group b Within borders of state A 1) Domestic State Terror 2) Domestic State Terrorism

Outside borders of state A 3) International State 4) State Sponsored Terrorism Terrorism

Table 2.1: Types of State-Terrorist Relations

21 In order to satisfy the criteria of state sponsorship of terrorism being specifically a relationship between the state and a terrorist group, the entire first column must be eliminated from the analysis. Indeed, the clear cases of domestic state terror (cell 1) as well as situations in which state officials, or agents of the state, perpetrate terrorism against dissidents or foreigners abroad (cell 3) were excluded when we narrowed our scope from the maximalist to the minimalist perspectives.

The same is true for the first row of the second column (cell 2) which represents situations in which the state facilitates, or sponsors, domestic terrorist groups that perpetrate attacks on their own civilians, including foreigners, within the boundaries of the state. Many cases of the three excluded cells have been considered state sponsorship of terrorism by scholars (see Hoffman 2006; Stohl & Lopez 1984, 1989; Levitt 2002;

Nacos 2010), yet either do not involve a terrorist group, or operate within the boundaries of the sponsoring state.

Thus, while a comprehensive examination of which types of state support qualifies as sponsorship remains, this study moves closer to a working definition by adopting the minimalist assumption that sponsorship is restricted to a relationship between a state and a non-state actor. Furthermore, to align with the understanding that state sponsorship is a foreign policy, it requires that the non-state actor must either reside outside the borders of the sponsoring state, perpetrate its attacks outside of those borders, or both (cell 4).

While this has set the broad parameters for discussing state sponsorship, the exact nature of the phenomenon is still unclear, illustrating why state sponsorship has been so 22 difficult to delineate conceptually. Indeed, Murphy’s argument that the “failure to distinguish carefully between terrorism and other kinds of violence makes “state support” and “state sponsorship” so broad in scope as to make the terms unmanageable from an operational perspective” remains a truism even today (1976, 31). Like terrorism in general, the heterogeneity of the actors, motivations, and manifestations has created obstacles to being able to define the phenomenon of state sponsorship clearly.

Despite this diversity there is also a common dynamic between the leadership of the sponsoring state and the nonstate terrorist group in which the leader provides the group access to the resources of the state in return for the group’s assistance in achieving a political objective. The resources provided to terrorist must be inherent to the sovereignty of the state, as opposed to those that are privately held, and are the final component in determining the definition. Thus, the task of the next section will be to determine the types of resources that should be included as state sponsorship.

INHERENT RESOURCES OF THE SOVEREIGN STATE States are unique in the international system in that they are the highest authority in the territory within the defined borders over which they have control and are expected to exercise the monopoly on the legitimate use of force within that territory.8 These features lead them to have a number of resources at their disposal that when transferred to terrorist groups, constitutes the state sponsorship of terrorism. This study will focus on seven categories of resources and argue that it is the intentional provision of these state

8 While Krasner (1994); Thompson (1994) and others have argued that control over these resources, and even sovereignty itself, is variable, this study assumes that unless the state is “failed” its leader has adequate access to, and reallocation authority of, these assets. 23 resources that qualify as state sponsored terrorism: protection (i.e. safe haven), access to territory, money, arms, intelligence and logistics, training, and diplomatic resources.

While states may overlook, acquiesce, or otherwise fail to prevent terrorists from using privately held resources, these actions fall under a different category than state sponsorship and should be examined as a separate phenomenon. The next section looks at each of these state held resources in turn.

Protection The provision of protection in the context of state sponsorship is understood to have occurred when the state9 uses its power and resources to protect terrorists from foreign governments without allocating them territory. This includes providing a group or its leadership safe haven, protection from extradition or prosecution, and allowing them to transit across territory. Protection is also a state providing cover for the group via front companies or allowing the group to open and maintain political offices.

The dynamics of safe haven support have only recently been explored in depth.

Byman (2005) argues that in the case of providing transit across territory, it is a very

“low-cost” type of assistance. Indeed, anecdotal evidence, and preliminary analysis of the data, indicates that providing protection is likely the most common form of support states provide to terrorist groups. It is still unclear, however, whether states that provide safe

9 Although this section addresses state resources as though the state is a unitary actor, it is merely for the sake of simplicity and clarity of the concept. Indeed, each of these could be further disaggregated to explore what aspects of the state structure and institutions control the resources and how they are allotted based on the division of domestic power. That however, is outside the scope of this project. 24 haven only to a terrorist group function under a different set of incentives and constraints than states that allow terrorists protection in combination with other forms of support.

Access to Territory The provision of access to territory occurs when the state allocates part of its territory for use by terrorists. It is important to note that this does not include instances where the state is unable to prevent terrorists from seizing the territory on their own.

Rather, it is the intentional provision of state territory to the group for use as training camps, sanctuary, or bases of operation. Sanctuary in this context differs from safe haven in that there is territory allotted specifically to the group rather than just allowing the group or leadership to seek refuge within the borders of the state.

Byman (2005) argues that for contemporary groups, this type of territorial sanctuary is the “most important” kind of assistance a state can give (54) and that the existence of these sanctuaries allows groups to “become far more potent” in that it

“facilitates all other forms of assistance” (65). However, the cost of sanctuary is expected to be higher for the state than some other types of support in that target states occasionally extract retribution for attacks, or invade to exterminate the threat, as Israel did against the PLO in Jordan in 1970 and in Lebanon in the late 1970s and early 1980s as well as for the group that is vulnerable to the state withdrawing support and even turning against it (Carter 2012).

25 Money The state can also provide terrorist groups with access to financial resources.

Although they are most likely to finance their operations through money laundering, drug running, kidnapping, extortion, bank robbing, and charity donations the financial assistance of states has historically been, and remains, a substantial operational boost for terrorist groups. This money can be used to buy weapons and explosives, create and maintain a logistics network, provide benefits to a terrorist’s family, and to provide recruitment incentives (Byman 2005). It also provides access to technology, such as cell or satellite phones, laptop computers, and the Internet, all of which have become indispensable to terrorist groups in recent years.

It is important to note that in the context of state sponsorship, the provision of money does not include instances where the state neglects to act to prevent private individuals from raising money or funding groups on their own, such as wealthy Saudi’s contributing to Hamas charities, nor does it include failure to prevent groups from fundraising, such as Sri Lankan groups associated with the Tamil Tigers in Canada.

Instead, the provision of money refers explicitly to act of the state itself providing limited or unlimited funds for the terrorist group whether or not it goes to finance a specific operation.

Arms One of the biggest material resources that modern states possess, and can provide to terrorist groups, is weaponry. This could range from light weaponry to heavy artillery, rockets, and bomb making materials. During the Cold War, access to conventional arms

26 was a key form of assistance that states provided to terrorist groups, yet the availability of arms on the contemporary international market, including black market, has somewhat decreased the value of this resource (Byman 2005). However, while light weapons, rocket propelled grenades (RPGs), and some surface-to-air missiles (SAMs) are easily purchased on the black market (Krause-Jackson & Walcott 2011),10 most heavier weapons, rockets, explosives, and technologically advanced weapons such as Hezbollah’s drones, are still provided by states.

Although providing weapons to terrorist groups changes in relative importance over time and helps to underline the variability of types of support in a sponsor-terrorist group relationship, it is also an easily identified and defined resource that is clearly lethal and in many cases, traceable. The provision of arms also includes instances where the state orchestrates or facilitates, indirect, third party arms transfers. However, this study excludes situations where the state merely does not regulate arms transfers on its territory.

Intelligence States also have access to highly sensitive foreign intelligence presumably gathered by, and intended for, state agencies. The category of intelligence is rather broad and can include information on topics such as troop movements, locations and identities of government officials, conventional military capabilities, and policy agendas. In the context of state support to terrorist groups, the category of intelligence encompasses a

10 See also the Small Arms Survey 2003 available for download at http://www.smallarmssurvey.org/publications/by-type/yearbook/small-arms-survey-2003.html accessed September 26, 2011 27 general knowledge transfer from the state to the group, whether this be advanced warning of counter-terrorism strikes, operational and logistical information regarding a target, or information about contacts within other governments or terrorist groups.11 The category of intelligence also includes the provision of logistical support, i.e. assistance by the state in moving personnel or equipment from one location to another.

Training States also normally contain institutions that are specifically geared toward training military and other personnel. In some cases, states will divert these resources toward training terrorist groups or allow members of the group to train alongside soldiers, security services, and other personnel of the state. In broad, general terms, training is understood to be the provision of organizational and operational “know-how” to the group. This skill-transfer can include basic and specialized military training, which can encompass everything from how to fire a gun to improvising explosive devices. It can also cover basic and specialized intelligence training as well as assistance in organizing and institutionalizing the group, and includes the provision of military advisors.

Although some argue that contemporary terrorism is trending away from official training camps and towards the use of the Internet for recruiting and training (Moghadam

2008), the basic operational training in weapons, explosives and tactics for members of terrorist groups remains, according to Byman, the “most common form of state assistance” (2005, 59). Training, regardless of the increased availability of information,

11 Transfer of knowledge can also occur in the other direction, that is, terrorist groups on the ground can provide intelligence or other information to the state. This occurrence is not coded as an instance of support, but it is an important component of the state-terrorist group relationship. 28 remains a predominately hands-on affair and is normally linked with other types of support such as the provision of territory and weapons.

Diplomatic resources The state also has significant diplomatic assets that can be made available to terrorist groups. This includes the provision of passports, visas, and other travel documents, such as when Libya provided passports to members of Abu Nidal who attacked the El Al ticket counter in Vienna in 1985 (Laquer & Alexander 1987). It can also include cases where terrorists are given diplomatic status as the Hungarians did for

Carlos the Jackal by allowing him to use the “diplomats only” passport control section at the Budapest International Airport (Yallop 1993, 489).12

It also includes either allowing terrorist groups to use the diplomatic pouch to move arms, or the state utilizing it on their behalf or allowing groups access to state embassies on foreign soil, such as when members of the Greek government offered PKK leader Ocalan safe haven in the Greek embassy in Kenya (USDS, “Patterns of Global

Terrorism 1998). While diplomatic support is a less common form of support, the implications for international relations is little understood.13

12 Note that the category of Diplomatic Assets ended up being rather localized to certain leaders and did not factor into some of the later descriptive analysis due to the very low number of cases. Rather, a category of “general support” was used to quantify situations where there was clearly support given, but the source did not specify the type. 13 This is another avenue in research that should be addressed more in depth in future studies on state sponsorship. 29 ARRIVING AT A DEFINITION The working definition being developed within the framework of the minimalist perspective begins by using the state-terrorist relationship as the foundation and then adding the understanding that state sponsorship is a type of foreign policy behavior.14

Finally, by including the requirement that the above state assets be intentionally provided to groups that perpetrate terrorist attacks, we arrive at the definition of state sponsorship that will be used throughout this study: state sponsored terrorism is the intentional allocation of particular state controlled resources to a nonstate group that perpetrates terrorist attacks in a foreign country.

Rather than being examined as a sub-type of terrorism (Yonah & Alexander 1985;

Wilkinson 1974; Schultz 1980; Anderson & Sloan 2002) or exclusively as a foreign policy designed to mobilize and respond to external threats (San Akca 2009; Bapat 2010) this study takes as its point of departure, the four assumptions stated above, that a leader’s primary objective is to retain power, that sponsorship is a rational decision, that the state is not a unitary actor and that leaders come to power in the shadow of the policy failure of the previous regime.

The next section outlines the many explanations for why states sponsor terrorism that are drawn from the international, domestic, and ideological realms of analysis. All fall short of providing a useful framework for understanding state sponsorship, while the

Theory of Leadership Survival, developed in Chapter 3, suggests a more unifying

14 Which the following section will explore more in depth. 30 structure from which to examine these profuse explanations for state sponsorship in the literature as well as in international relations.

LITERATURE REVIEW The existing literature on state sponsorship supports the notion, advocated here, that leaders support terrorist groups when faced with critical threats to their regime. There is little agreement, however, as to the origin or nature of these threats. The literature on state sponsorship tends to fall into the same three categories that IR scholars emphasize, first, those that give pride of place to international or strategic variables; second, those that prioritize domestic level explanations; and finally, scholars that view ideology/ideas and norms as the key to the analysis. Yet, rather than being set up as competing theories, scholars of state sponsorship use these levels of analysis in loose combination. In other words, the incentives and constraints that originate in the international sphere seem to line up with the domestic and ideological variables to produce a policy of state sponsorship (see especially Byman 2005; San Akca 2009).

Causal theories of state sponsorship are generally underspecified, normative, and tend to “throw in the kitchen sink.” Indeed, one striking aspect of work on state sponsorship is the frequency with which theories that utilize strategic variables such as state size, military capabilities, and external threat environment (Bapat 2007; Levitt 2002;

Livingstone & Arnold 1986: 15; Nacos 2010; San Akca 2009; see also Hoffman 2006) also incorporate explanations that focus on domestic level variables such as regime type, concerns for legitimacy, and domestic political “spillover” (Cline & Alexander 1986, 31 Moore 1997; Livingstone & Arnold 1986) without attempting to disaggregate causal mechanisms or develop clear theoretical arguments. Moreover, while not discounting any of the above explanations, state sponsorship is also causally attributed to a leader’s zealous ideological drive at the expense of rational power politics (Moore 1997). In other words, while it is widely accepted in the state sponsorship literature that there is an interaction between international and domestic level variables, the way in which these variables interact in order to produce the outcome of state sponsored terrorism has been inadequately specified and rarely subjected to empirical testing.

Noting these overlapping pressures, San Akca (2009) is the first to develop a theory of state sponsorship of nonstate-armed groups that intentionally focuses on the combination of the strategic interests, ideational affinities, and domestic incentives. She concludes that sponsorship occurs when the state is too domestically (largely ethnically) fractured to internally balance, nor can they muster allies to externally balance the existing external threats to the state. In this case, nonstate-armed groups function as a substitute for traditional allies (see also Vinci 2009). Ideational affinity between the sponsoring state and the nonstate-armed group is shown to be a statistically significant variable, yet, it remains unclear what role ideology plays outside of facilitating cooperation. While this study uses San Akca’s work as a point of departure and accepts many tenants of her theory, most importantly the assumption that traditional external balancing through allies is not always an attractive or possible option for state sponsors, her focus remains primarily on the logic of balancing vs. the logic of retaliation in

32 international relations and does not pay adequate attention to how domestic politics shapes a leader’s incentives to sponsor terrorism.15

Many works in the field of international relations and foreign policy formulation, especially that of grand strategy, have begun to look at how international and domestic variables interact to influence policy formation with the state executive16 largely serving as the point of integration (Dueck 2006; Zakaria 1997; Trubowitz 2010). This study, which accepts the basic premise that state sponsorship is a type of foreign policy, builds on the theoretical foundation of integrating international and domestic variables at the level of the state leader and applies it to the study of state sponsored terrorism.

While attention to the effect that nonstate actors, such as terrorist groups, have on international politics and the use of these groups strategically by states is already increasing among IR scholars, many of these works, pull solely from IR literature on civil war, ethnic conflict, and rebel groups (Salahyan 2007 & 2008; Bapat 2006, 2007 & 2011;

Vinci 2009) and have not incorporated the full range of literature on state sponsorship.

Where work specifically on state sponsorship has been integrated, it has not been subjected empirical testing (some prominent examples are Jenkins 1974; Mickolus 1989;

Byman 2005). The task of this chapter, then, is to examine where the literature on state sponsorship aligns with, and can inform, the work of IR scholars addressing states that support terrorist groups.

15 These shortcomings will be explained in depth more in Chapter 4 when an in-depth analysis of her research design, data collection, and theoretical assumptions are made. 16 This study adopts this technique. While some studies focus on the leader that is largely responsible for formulating foreign policy, i.e., the Foreign Policy Executive (FPE) (Lobell 2009) and others define this executive more broadly as regime, in the interest of parsimony and generalizability across cases, this study adopts the term “leader” to describe the state executive as identified by Bueno de Mesquita et. al. (2003). 33 This remainder of this chapter will be organized as follows. First, it will examine existing theories of state sponsorship that focus on external threat to the state, looking specifically at explanations that hinge on state strength, the role of allies, and costly signaling. Second, it will explore the literature that addresses the domestic level and causal arguments related to regime survival, paying special attention to regime type, legitimacy, and regime cohesion. Third, the same will be done with theories that give causal weight to ideology. Fourth, the literature that has made attempts to develop theories that integrate the international, domestic, and occasionally ideological levels will be explored. Finally, from these literatures, a series of general premises will be extracted, which in Chapter 3 will be refined into testable hypotheses. In order to develop a solid empirical foundation for the study of state sponsorship; I will empirically test my hypotheses though large-N analysis using a dataset constructed specifically for this project.

EXTERNAL THREAT External threats and territorial conflicts are regular features of international politics and have played a prominent role in the explanations for why states sponsor terrorism. The notion that using terrorist groups affords a state significant strategic advantage vis-à-vis their rivals is a common theme in nearly all works on state sponsorship (Mohindra 1993; Byman 2005; Nacos 2010; Cline & Alexander 1986;

Hoffman 2006). Even those studies that give a heavy causal weight to radical ideology acknowledge that strategic political considerations play a key role (Shay 2005).

Strategically, terrorist groups can be used by the state as force multipliers, as substitutes 34 for conventional methods of destabilization (Livingstone & Arnold 1986), and are an effective way to project power beyond one’s borders (Byman 2005). Indeed, both Byman

(2005) and Cline & Alexander (1986) argue that for many states the primary motivation for state sponsorship arises from a desire to destabilize or overthrow neighboring regimes, shape the domestic opposition within them, and take advantage of the position this affords the state in intra-regional politics.17

There are several logics that see external threats as the cause of state sponsorship: first, scholars have argued that support for terrorism is a straightforward strategic tool of weak states unable to compete in international conflicts with weak conventional militaries (Hoffman 2006); second, it is an adaptive technique used by states unable to muster internal resources or secure external allies (San Akca 2009); and third, that it serves a costly signaling function that shows the credibility of commitment to a particular policy preference (Bapat 2007). These explanations are all strategic in nature and while some have the causal mechanisms more teased out than others, all point to a leader who makes international norm-defying foreign policy choices under severely constrained circumstances.

Weak States? The conventional wisdom is that weakness causes states to engage in the use of terrorist group “proxies” because they are too weak to achieve their foreign policy goals through conventional means (Livingstone & Arnold 1986; Levitt 2002). Indeed, Hoffman

(2006) attributes state sponsorship specifically to state weakness and inability to muster

17 The effect that this strategic advantage affords the leader domestically, however, is not explored. 35 the conventional resources necessary to project power and protect state interests in the international system. State sponsorship, then, is largely understood to be “a tool of the weak to use when non-terrorist forms of influence are insufficient or unavailable” (Volgy et al. 1997, 211). Importantly, in this picture, like most dealing with warfare, the leader and the state are assumed to be synonymous and the military a measure of capabilities and an instrument of policy, albeit a weak one in this case.

Some scholars, however, have challenged the premise that weak states are most likely to sponsor terrorism in the face of an external threat and have drawn on the myriad examples of sponsorship by strong, centralized states such as the Soviet Union, Libya,

North Korea, and Iran (Nacos 2010). Still others claim that both weak and strong states can sponsor terrorist groups. For example, Byman (2005) develops a scale showing that state capacity varies and has a key impact on sponsorship. He suggests that weak states are more likely to be passive sponsors while strong states are more likely to be active sponsors. Still others point to the prevalence of terrorists using weak and failed states such as Lebanon, Afghanistan, and Somalia as safe havens, or bases of operations and staging grounds, for attacks (Lesser 1999; Rotberg 2002; Patrick 2006; and Rice 2008).

San Akca’s (2009) work was the first to subject state strength to empirical testing, outlining a rationale for both strong state and weak state likelihoods for sponsorship that drew from IR literature. She argues that relatively weaker states are most likely to sponsor non state armed groups (NAGs) because of the logic of balancing (i.e., states will increase their capabilities to balance stronger states however they can), while relatively

36 stronger18 states are likely to sponsor because of the logic of retaliation (specifically, weaker states can not retaliate). She finds that contrary to much of the literature on state sponsored terrorism, which argues weak states are the most likely sponsors; it is in fact the relatively stronger states (vis-à-vis the target state) that are more likely to sponsor

NAGs.19 She concludes that sponsorship is not a balancing behavior per se, but rather is representative of a situation where one state is taking advantage of an opportunity to strike another via a proxy and with little fear of retaliation from the target state.

Subsequent empirical examinations concurred in part, finding that states need to be strong enough to credibly commit that they will get the terrorist group under control if they get the concessions they want from the target state, yet they also have to be weak enough that the concessions from the target state have to be tremendous in order to warrant the increased resources that the sponsoring state requires to bring the groups down (Bapat 2011).

The strength of the state, then, does play a role in state sponsorship, yet not in the way that most scholars have thought. The Theory of Leadership Survival, developed in the next chapter, will argue that this perceived state “weakness” is not an empirical fact, but rather is a function of the inefficiency of the army, which is kept intentionally weak in order to preserve the leader’s position of power. Furthermore, far from being a way to enhance the capabilities of the military, this study will show that terrorist groups are used

18 Her study measures state strength using the Correlates of War National Capabilities data, which is a composite variable of a country’s energy consumption, iron & steel production, military personnel, military expenditure, total population, and urban population. 19 The relevance for these findings to sponsorship of terrorist groups in particular is unclear and will be subject to empirical testing in Chapter 3. 37 explicitly to circumvent the military, enhance the capabilities of competing institutions

(specifically the intelligence services), and allow the leader to maintain power in the face of converging internal and external threat to his or her rule.

Isolated States with Few Resources The second logic connecting external threats with state sponsorship in the literature was further by developed by San Akca (2009) who argued that while all states have a limited number of alliance choices (Morrow 1991; Lassinantti 2001) some are simply unable to muster external state allies. When this is combined with the inability to balance internally (extract from the domestic population), because of obstacles to internal mobilization, states turn to terrorist groups who serve the same function as external state allies and allow the sponsoring state to stay in the fight.20

However, San Akca’s study is still unable to account for why, given their apparent dearth of resources, “domestic trouble,” and lack of conventional strength, the state would choose to invest resources and unconventionally balance the threat rather than appease (grant concessions) to buy time or to settle the conflict (Ripsman & Levy

2008) nor does she give a clear explanation as to why states are unable or unwilling to enter into traditional alliances. Her point remains that is not necessarily that leaders prefer to sponsor terrorist groups over other “security on the cheap” alternatives, but rather that there are no alternatives. However, I argue in Chapter 3 below that even when available, the use of traditional state allies is not without tremendous risks of its own. While one of

20 The isolation argument is based partially on the premise that non-state armed groups are able to preserve the international balance of power in much the same way states do (Vinci 2009). 38 the key dangers of alliances are that states can be “entangled” or otherwise dragged into an interstate confrontation that they do not desire, a second is that alliances can dramatically increase the operational and organizational capacity of the military and highlight existing policy divergence between the military and the leader. For leaders already facing a threat from their militaries in the form of a coup d’état, traditional (state) alliances can increase this threat to unacceptable levels and trigger internal and onmi- balancing mechanisms in response.21

The notion that terrorist groups are a direct substitute for traditional state

“alliances” is relatively new, yet most work on state sponsorship assumes an external threat that is being met, in part, by delegating the task to terrorist groups, whether conceptualized as sub state “proxies” or as “agents” in the context of principal-agent analysis (Byman & Kreps 2010). These groups are thought to be able to contribute a number of advantages to the sponsoring state beyond mere power projection, including offering a specialization in unconventional tactics, increasing the credibility of the sponsoring state, and entrenching the sponsoring states’ policy preferences (Byman &

Kreps 2010). In this way, terrorist groups do function roughly as a type of alliance that also send costly signals to rival states, yet it is still unclear why a state, or state leader as the actual case may be, would want to invest in this type of “costly” signaling when other

21 We will explore this dynamic in the section on Saddam Hussein’s Iraq in Chapter 5. 39 forms of conflict resolution and “security on the cheap” are available.22 That is unless any other type could cost the leader his job.

Costly Signaling The argument that terrorist groups provide the state with a level of specialization in unconventional or covert warfare is widespread, but the idea that sponsoring terrorist groups can also increase the credibility of the sponsoring state is relatively new (Bapat

2007; Byman & Kreps 2010). Bapat (2007) conceptualizes sponsorship as a form of costly signaling that shows the sponsoring state has a tremendously high level of commitment to a policy position. State sponsorship, he argues, has the effect of both increasing the chances of bargaining failure between the sponsor and target state—more likely to lead to war— while pushing the target state, via terrorist attacks, to agree to concessions that are so enormous that the sponsoring state has the incentive to crack down on the militants. In other words, although an agreement between the sponsoring state and the target state may be harder to come by than if terrorist groups were not involved, the agreement that is eventually arrived at will be more likely to be favorable to the sponsoring state.

For many scholars, then, state sponsorship is a story of the state intending the terrorist group to be leverage against an external threat with “the ultimate objective of forcing the target state into submitting to the will of the aggressor state” (Ceroli 1994,

22 Again, this “costly signaling” model assumes that the intended audience is the target or third party state rather than the leader’s ruling coalition. 40 67). From the abundant literature on state sponsorship we can derive our first premises23 that will guide in the formulation of the Theory of Leadership Survival. First, leaders that face an external threat are more likely to sponsor terrorist groups. Second, This choice is informed in part by the availability and distribution of internal resources. Third, the strength, efficiency, and ability to mobilize the national military are also factors in state sponsorship.

Whether used in the context of coercive diplomacy or violent coercion, the studies that focus on external/strategic explanations all point to a policy of “security on the cheap” that is developed in order to extract concessions from adversaries and increase the sponsoring states’ relative power. While these effects of state sponsorship may be accurate, these explanations often conflate outcomes (results) with motivations

(intention) and cannot actually account for the variation of the policy choice in support of terrorism within the international system. What, then, differentiates states that use terrorist groups from those that do not? If the conventional strength of the state and the external security environment alone cannot make this distinction, we have to look to additional levels of analysis.

DOMESTIC VARIABLES A second, complementary, strand in the literature approaches state sponsorship as an outcome of the domestic strategic environment. Regime type (Cline & Alexander

23 These will be framed as “premises” rather than “hypotheses” because they themselves will not be tested. Rather, they are being isolated as general axioms that scholars in IR and state sponsorship can agree on. In the theory section below, these premises will be teased out in the context of the theory, refined into hypotheses, and subject to empirical testing. 41 1989), the legitimacy of the regime (Nacos 2010), and the challenges to, and cohesiveness of, the regime (San Akca 2009) play major roles in the domestic level explanations state sponsorship. This mirrors the strand of the IR literature and the domestic theories of foreign policy that tend to focus on political struggles and coalitions within the regime (Allison 1971; Snyder 1991) and struggles between the regime and other influential societal actors in resource allocation and policy formulation (Christensen

1996; Fischer 1975; Olson 1982). This section will lay out the explanations for state sponsorship that have been attributed to regime type—including rogue regimes, legitimacy, and regime cohesion.

Regime type The conventional wisdom regarding regime type is that state sponsorship is most likely used as a foreign policy tool against democracies by authoritarian and totalitarian states (Cline & Alexander 1986; Crenshaw 1981; Kegley 1990; Livingstone & Arnold

1986; Wilkinson 1977; see also the “radical regimes” of Moore 1997). State sponsorship is largely thought to be a spillover from repressive domestic political practices into foreign policy (Cline & Alexander 1979; see also Shay 2005; Levitt 2002). Specifically, it is thought that since authoritarian and totalitarian states use domestic tools of terror, intimidation, and coercion against their domestic political opposition (Arendt 1951) and

“states duplicate patterns of domestic politics in the international arena and apply the same norms of conflict resolution in both domestic and international settings” (Caprioli &

Trumbore 2003, 45) then these states are most likely to sponsor terrorism.

42 The second argument linking state sponsorship to regime type draws on the association in the literature, and in policy circles, between state sponsored terrorism, authoritarian regimes, and rogue states—those that place “a high priority on subverting other states and sponsoring non-conventional types of violence against them” (Rubin

1999, 1) including terrorism. In fact, Rubin argues, “it is virtually inevitable that a state considered a rogue will be a repressive dictatorship” (1999, 2). This is more than an association, then, it is an equation.

Other studies absorb rogue states and other “international deviants” under the normative rubric of “renegade regimes.” The development of WMD, external aggression, support for terrorism, and “externally harmful domestic repression” are the defining criterion of these renegade regimes, nearly all of them authoritarian (Nincic 2005). While

Caprioli & Trumbore’s (2003) study tests the relationship between state sponsorship and rogue states, they do not take regime type explicitly into account. Rather, they argue that state sponsorship can be predicted by a state’s “Rogue State Index” (RSI) score of 0

(low) through 8 (high), which is a composite score of several variables: the level of gender equality, ethnic discrimination, and domestic political repression. They find that for every increase in the RSI, there is a 63% increase in the likelihood of a state sponsoring terrorism. However, given that the average composite scores of Cuba (2.5),

North Korea (3.5), and Libya (3.5) are all lower than the UK (3.6) and France (4.5) while

43 the USSR-Russia scores on par with the UK at a 3.6, it is unclear that this metric either valid or useful.24

Other works on state sponsorship place the blame on the domestic institutional features of authoritarian states such as the lack of democratic constraints on decision- making and resource allotment (see San Akca 2009; Schultz 2001), while still others argue that the clandestine nature of state sponsorship is likely to be more successful in states that lack a free media (Byman 2005). The relationship between regime type and state sponsorship will be explored more in Chapter 3, yet the high correlation in the literature between regime type and state sponsorship leads this study to an additional premise that authoritarian regimes are more likely to sponsor terrorist groups, though the

Theory of Leadership Survival will demonstrate that this association is not due to these reasons widely assumed in the literature on state sponsorship, but rather because the structural make up of a small winning coalition regime (Buena de Mesquita et al. 2003) predisposes the leader of authoritarian regimes to particular internal threats.

Legitimacy A second domestic variable is the leader’s--normally explored as the regime’s-- legitimacy. There are several arguments that have been put forward linking problems of regime legitimacy to state sponsorship. First, it is thought that the leader uses the support of an outside terrorist group specifically to “divert domestic radical’s frustrations to outside targets” (Nacos 2010, 118). A second, though closely related, logic parallels the

24 Indeed, the weighing in of Iran at an average score of 7.1 may have skewed the analysis significantly, given that the data years were 1980-2001. 44 diversionary war literature. These explanations understand that when a state is weak internally and the leader is threatened by domestic groups, then backing an external terrorist group could increase regional tensions between the state and its neighbors and may have the effect of allowing the leader to rally the population around the regime and

“reinforce a state’s influence at home” (Byman & Kreps 2010, 4). While the origins of this weak legitimacy are not adequately explored, nor the main contenders for power identified in the literature on state sponsorship, these reoccurring themes lead to an additional premise that regimes with weak legitimacy are more likely to pursue strategies of “security on the cheap” by sponsoring terrorism.25

Regime Cohesion The last piece of the domestic puzzle, regime cohesion, is connected to both the weak state and the legitimacy arguments discussed in the sections above. Rather than focusing on capabilities and relative power, this argument traces sponsorship to a weak and decentralized state with fractured institutions, or regime factions, some of which may even be funding terrorism at the expense of the other parts of the government (Byman

2008).

There are two possible motivations for sponsorship within this framework. First, an institution or branch of the government can use support for terrorism to bolster or entrench its power vis-a-vis other institutions, such as when the Pakistani intelligence services (ISI) provide support to the same groups in Afghanistan that the civilian

25 Importantly, authoritarian regimes are not the only ones that can face legitimacy problems, although they may be the most prone.

45 government has vowed to fight and which the military forces are actually engaging in battle. The second motivation is simply for personal gain, whether it is Gambian military officers selling arms to the Movement of Democratic Forces of Casamance or Togo’s president Gnassingbe Eyadema using Togo’s end use licenses to provide arms to UNTIA for a substantial amount of money (UNSC S/2000/203 May 7, 1999 http://www.un.org/News/dh/latest/angolareport_eng.htm).26 The Theory of Leadership

Survival, developed in Chapter 3, will explore how regime cohesion is, in fact, a major factor in state sponsorship. Especially in cases where there is extreme policy divergence between the military and the state leader, and an existing threat of military coup, I argue that leaders do, in fact, cultivate certain institutions (intelligence services) at the expense of others (the military) and that the sponsorship of terrorist groups provides the leader an avenue to both balance the threat emanating from the military and to counter external threats.

In sum, domestic level explanations have long been under-examined in the study of state sponsorship despite the fact that these factors can “lead a regime to support terrorism, even when it is not in their state’s strategic interests and when the elites’ ideological convictions are thin at best” (Byman 2005, 47). Indeed, the Theory of

Leadership Survival will show that these domestic level explanations are necessary in order to understand state sponsorship. Importantly, the domestic weakness of the leader, vis-a-vis other domestic actors—most importantly the military--interacts with strategic

26 Which upon closer look may also be related to regime survival, since in many small coalition patronage systems the ability to buy-off other elites is key to maintaining power. 46 level interests and threats in order to sufficiently incentivize the sponsorship of terrorist groups.

IDEOLOGY The last, and very prominent sets of explanations scholars use to account for the decision to sponsor terrorist groups revolve around ideology. The question, however, is really regarding the causal mechanism of ideology, that is, how exactly it functions as an explanatory variable in the puzzle of state sponsorship. Why would we expect a state that has an ideological similarity with a terrorist group to be more likely to sponsor them? The answer has been framed in three major ways: first, that there is a natural alliance between the terrorist group and state, which is often accompanied by a common enemy, goals, or grievances (Nacos 2010; Shay 2005; Byman 2005; Hoffman 2006; Cline & Alexander

1986; Levitt 2002). Second, that state sponsorship is a medium to expand the power and influence of the state though ideological diffusion (Slann & Schechterman 1987; Caprioli

& Trumbore 2007).27 Finally, that providing support to terrorist groups is a way to lower transaction costs and lessen the “agency slack” that occurs when terrorist groups seek to put their own preferences ahead of their state sponsors (Byman & Kreps 2010). Each of these pathways will be explored in turn.

27 Ideology can also function as a way for states seeking strategic advantage to “enhance international prestige” or to “export (its) political system” (Byman 2005, 43 & 41). In fact, in his 2005 study, Byman argues that ideological beliefs were a leading concern for the sponsoring states in seventeen cases (18%), while they contributed to 26% of the 95 post-Cold War cases. The dominant role of ideology for these sponsors was a “desire to export” both their ideology and political system, which is an issue of strategic advantage as well (45). 47 Natural Alliance Even within the context of the strategic and domestic motivations mentioned above, some studies find that “in several prominent instances of state support for terrorism the ideological driver behind the relationship cannot be overstated” (Byman &

Kreps 2010, 5). Indeed, San Akca’s (2009) study indicates that states are more likely to sponsor a non-state armed group with which it shares an ideological affinity. This margin of support, however, is rather slim. While 40% of the states in her study shared ideational ties with the non-state armed groups they support, a full 37% of the states supported non- state armed groups that were not ideationally affiliated. Again, the question arises of what exactly the role is that ideology plays in state sponsorship.

This study incorporates the widespread notion that indicates ideology is relevant in the study of state support for terrorism, yet does not accept it as a causal factor.28

Rather, ideology serves as a means to facilitate cooperation (Byman & Kreps 2010),29 signaling competency and fidelity to an often highly ideological ruling coalition. In highly ideological regimes, such as Iran, support to terrorist groups is expected to extend specifically to ideologically affiliated groups. Indeed, in Iran, the “combination of radical ideology and religion alongside political considerations, …turn the use of terror within the Iranian political framework into a rational and inevitable component” (Shay 2005, 19; emphasis mine). Yet, there are some indications that this ideology-support relationship is variable rather than constant. Although Iran has supported Muslim groups since the

28 Ideology cannot account for the decision of state sponsorship. It can, however, assist the leader in determining which groups will be supported. Thus, it does not play a causal role, but merely helps shape what that relationship looks like once the decision has already been made. 29 This will be explored in greater detail in the section on lowering transaction costs below. 48 revolution in 1979 and continues to support groups such as the Palestinian Islamic Jihad

(PIJ), Hamas, and Hezbollah, it also provided support to multiple Kurdish nationalist groups in Northern Iraq none of which shared Khomeini’s ideology, but rather had converging strategic goals. Iran, although having lost some of its revolutionary zeal by the time of the Chechen wars, denied Iranian assistance to the ideologically affiliated

Islamic Chechen separatist rebels in Russia, despite sharing a border, based solely on the fact that Iran’s strategic relationship with the Russian government proved far more important (Malek 2008).30

Some scholars, however, discount the strategic angle completely and argue that ideology can become the end in itself; that the state can become “fundamentalist,” thus leading the state to violently disregard domestic and international law in favor of introducing a higher order (Slann & Schechterman 1987; see also Caprioli & Trumbore

2007). Moore (1997) concurs and argues that there is a “radical regime syndrome,” which has the state sponsorship of terrorist groups as one of its core characteristics. The state, from this perspective, sponsors terrorism not out of national interest, but rather is driven pathologically to challenge the existing international order to realize its utopian vision.

30 Some such as San Akca (2009) understand these relationships to be “substitutes” for alliances with other states. In that vein, it is worth an examination of Steven Walt’s (1987) ideological alliances between states. He argues that these might be a way for states to advance and defend their own political principles, lessen distrust and alleviate fear, or enhance legitimacy in weak regimes by aligning to a larger movement popular among the people (34-35). Additionally, he notes that in general, the more similar the ideology of states, the more likely they are to ally and also that the “impact of ideology on the choice of alliance partners will be exaggerated” that is, “statesmen will overestimate the degree of ideological agreement among both their allies and their adversaries” (40). In sum, this indicates that ideological affinity is neither a puzzle nor an absolute, but is, as argued above, merely a feature of the sponsor-terrorist group relationship that lowers transaction costs and facilitates cooperation. 49 Yet, ideology may merely provide a way for the state to rationalize its behavior (Leavitt

2002) or as this study argues, facilitate the achievement of its interests.

Lower Transaction Costs Another way that ideology plays a role in state sponsorship according to the literature is as a low-cost, highly effective way to the reduce transaction costs and agency slack, and to improve monitoring of the tasks that the state delegates to terrorist group

(Byman & Kreps 2010). Ideology provides a clear, and mutually agreed upon, bottom line for behavior and goals. States and terrorist groups operating within this framework can both advance more efficiently towards both their individual and mutually stated goals. Ideology, then according to this view functions as a tool to facilitate the smooth transmission of ideas, values, and normative standards of behavior.

Anecdotal accounts of strong relationships between ideologically oriented regimes and the terrorist groups they support abound in the literature and the empirical findings of some preliminary works on this relationship seem to support this notion

(Byman 2005; San Akca 2009). Yet, how this ideological affinity results in a state sponsorship is best understood as a low-cost, high efficiency tool for cooperation between the terrorist group and the sponsoring state. Indeed, ideological affinity is a common attribute of the sponsor-terrorist group relationship, rather than a causal explanation for why a state would opt to adopt the policy option of sponsoring terrorist groups.

Calculations of strategic advantage, external threat, and concerns over domestic politics are often intertwined with ideology in explanations of state sponsorship in the 50 literature. In order to get a sense of how these interact, next this study will examine the literature that has converged these levels of analysis in an attempt to get a more realistic and comprehensive understanding of state sponsorship. It is also this convergence that provides a true jumping off point for the development of the Theory of Leadership

Survival.

CONVERGENCE OF THE VARIABLES Regardless of whether one is looking at strategic, domestic, or ideological explanations for why states sponsor terrorism, there are a number of questions that remain unanswered: why would states choose to sponsor terrorist groups rather than to achieve their strategic goal through other avenues? Why, for example, if the state is weaker than its regional rivals and faces disadvantages with their conventional forces, do they not channel their resources to build conventional armies, work through the United

Nations or other multilateral institutions, accept the status quo, appease the stronger state and make concessions, or form strategic alliances with other states? I argue that the above explanations for state sponsorship are not equipped to address these questions and in

Chapter 3, I offer my own argument as to why leaders opt for state sponsorship rather than an alternative strategy for “security on the cheap.” First, however, it is worth examining some foundational work that does address these questions, but has as its point of departure, the state facing an external threat for which it can neither internally nor externally balance. Again, I adopt a similar point of departure, but rather than accept this

51 failure to balance at face value, I problematize the inability to muster internal resources or adopt external or other “security on the cheap” strategies.

San Akca’s (2009) recent work on the state sponsorship of non-state armed groups (NAGs) combines the three incentives covered in detail above of strategic interest, ideational factors, and domestic instability to develop a leader’s selection model of state sponsorship. Her model outlines several circumstances in which a leader would be more likely to make a decision to provide support for a NAG based on 1) whether the

NAG targets an enemy state; 2) whether the NAG is “ideationally contiguous” with the state; and 3) whether the leaders “face risk to survival at home, which imposes constraints on their ability to extract and mobilize resources to deal with an external adversary” (29). Thus, she argues, states that face external enemies, but are unable to mobilize domestic resources or secure external allies to meet the threat are most likely to sponsor non state armed groups.

There are three major contributions made by her study that I adopt in my own work. First, examining these variables in combination rather than in isolation allows for a far more accurate picture of the policy formulation process and the interaction of external and internal pressures affecting decision makers (Maoz 1990; Putnam 1988). Second, the theory is generally consistent with the existent literature on state sponsorship and brings it squarely into the realm of international relations. Third, it correctly emphasizes the importance of the survival risks to leadership within the domestic arena as a variable.

The Theory of Leadership Survival developed in Chapter 3 draws on many of these insights, yet shows the leader’s survival risk, understood as combined internal and 52 external threat, not as a problem of internal balancing (extracting and mobilizing resources) and external balancing (allying with other states), but as a problem of omnibalancing (David 1991). Omnibalancing, i.e., a situation in which a leader must balance against both internal and external threats, was developed to explain state balancing behavior in much of the Third World that appeared suboptimal or counterintuitive to scholars who assumed the origin of all threats were external. It draws on the insight that leaders sometimes face threats to their power that are different than the threats faced by states and thus will sometimes make strategic choices to protect their rule that are in contrast to the rational interests of the state. For many leaders, the primary threat does not come from external states, but rather from internal groups that are jockeying for power within the state.

Several pieces of the puzzle can be found in the literature on state sponsorship, though this dissertation is the first to bring the causal mechanisms and a unifying framework together by centering on the dual pressures the leader faces from external and internal threats simultaneously. In the existing literature, there are indications that leaders who face an external threat are more likely to sponsor terrorist groups against that threat, pulling state sponsorship, then, into the realm of foreign policy and strategic assessment.

There is also an ongoing dispute regarding the effect state strength (relative power, capabilities, etc.), or lack thereof, has on the policy choice to sponsor terrorism, which my study will show has been defined incorrectly because in certain circumstances an increase in capabilities can actually increase the threat posed to the leader by the military and thus the likelihood that leader will sponsor terrorism. 53 Second, anecdotally at least, authoritarian regimes appear to be more likely to provide support to terrorists. Although whether this is because dictators are more likely to use repressive tactics externally as well as internally or because the leader enjoys freedom from democratic institutional constraints remains unclearly specified in the literature. My study, again, will examine what exactly it is about authoritarian regimes, using insights from Bueno de Mesquita et al.’s (2003) selectorate theory that makes them more likely to support terrorism.

Third, the literature returns time and time again to the notion that regimes that are decentralized, fractured, and/or have weak legitimacy also appear to be more likely to support terrorism. My theory, as outlined in the next chapter, will argue that this competition within the domestic realm itself, specifically between the military and the leader, plays a primary role in the state sponsorship of terrorism.

Thus, taking the existing explanations for state sponsorship that have been so prominent, but under examined, in the literature and looking at them in combination and within a framework of an omni-threatened leader, rather than as isolated factors influencing the “state” provides a more accurate account of the policy decision leaders make to support terrorist groups.

The task ahead then, is to take these existing pieces of the puzzle and develop this unifying theoretical framework of state sponsorship as a tool of leadership survival that incorporates both the external and internal threat environment of the leader and adds the additional variable of the leader’s relationship to the organization most capable of removing the leader from power: the military. It also will clearly specify the role that 54 terrorist groups, and the intelligence services that arm and equip them, play in helping to alleviate these internal and external threats, i.e., provide a way for the leader to strike at

(balance) the external threat while preventing the military from gaining the organizational and operational capabilities to threaten her rule. In other words, terrorist groups, and the intelligence services that arm and equip them, are some of the most effective omnibalancers available to threatened leaders.

55 Chapter 3: The Theory of Leadership Survival

The literature on state sponsorship consistently highlights a situation where a ruler, facing external threats and mobilizing for external conflicts, provides support for a terrorist group to increase their chances of victory, or at the very least, to help prevent defeat. Studies also show that internal pressures, weak legitimacy, and frictions or splits between the ruler and other internal actors compel a leader to sponsor terrorist groups.

This study attempts to place these observations within a single framework that understands a decision to sponsor terrorism as a rational response for leaders that face these threats simultaneously.

The Theory of Leadership Survival, developed below, accepts that in some cases, state sponsorship occurs when the leader appears to face only an external threat.31 Yet, the most intractable and intense cases tend to emerge when leaders face simultaneous internal and external threats to their rule, especially when these internal threats come from the military itself in the form of coup d’état.32 In these cases, mobilization for war, regardless of the reason, requires that more resources be diverted to the military. As the military grows in organizational strength and capability, the internal threat to the leadership from the military is also increased. This is a situation that external aid can

31 What will be specified as a High External & Low Internal (HELI) threat environment. Further research into this question, however, is expected to reveal that there is a level of policy divergence between the military and the leader that causes the leader to circumvent the military and empower intelligence services that utilize terrorist groups to balance the military, secure policy goals, and maintain the leader’s position of power within the regime. 32 When segments of the leader’s own population (i.e. ethnic, insurgent, or terrorist groups) also mount a challenge to the regime, this dual security threat can be heightened even more. 56 make even worse.33 The leader then is faced with a dilemma, if she does not mobilize the military forces for war or otherwise respond to the external threat then she risks losing power within the ruling coalition on the basis of perceived weakness, especially if the crisis relates to one of the pillars upon which she has established her legitimacy, and potentially risks even the survival of the state. However, if she does mobilize the military for war then she risks creating a force that has increased organization, capabilities, and investment in their perception of security (which does not always line up with that of the leader). This greatly ups the potential for the leader to be overthrown by coup d’état, especially in cases where a coup threat already exists.34 By sponsoring terrorism, and thus investing some of the state’s military and power-projection capabilities outside of the military’s control and into the intelligence services and supported terrorist groups, the leader is able to counter the external threat, minimize the internal one posed by the military, and signal to the other members of the ruling coalition that everything is under control.

I aim to explain how the combination of internal and external threats, when filtered through the relationship that the leader has with the military, compels a leader to chose state sponsorship of terrorist groups in order to maintain power. This theory centers on four foundational assumptions. First, that a leader’s primary objective is to retain power (Bueno de Mesquita et al. 2003; David 1991; Downs 1957; Nincic 2005;

Trubowitz 2010); second, the decision to sponsor a terrorist group is strategic and rational

33 As the case examining Iraq during the Iran-Iraq war in Chapter 5 will show. 34 Even in situations where the leader faces low threat of coup, when her legitimacy hinges on a foreign policy objective and the military is an obstacle, she will circumvent the military to achieve that goal. 57 (Byman 2005; Hoffman 2006); third, the state is not a unitary actor, but rather the leader sometimes faces acute threats separate from those the state faces (David 1991); and fourth, all new leaders come to power with their legitimacy hinging on their ability to succeed where their predecessors failed on key policy issues. When these are added to the premises isolated from the state sponsorship literature a new theory begins to emerge.

A KEY MISSING VARIABLE: THE MILITARY International Relations scholars tend to assume that the state’s military is subsumed under the command of the unitary state. In fact, the military, when considered at all, is merely a component variable of a state’s capabilities rather than an actor itself

(Mearshimer 2001; Van Evera 1999; Walt 1987; Waltz 1979). Studies that explain state sponsorship as arising from the circumstance of a weak conventional military likewise assume that this military is wielded as a willing tool of the state (Byman 2005; Cline &

Alexandar 1989; Hoffman 2006). Yet, the civil-military relations literature clearly shows that the submission of a well-armed national defense force to the civilian leadership of a state is variable, not constant (Desch 1999; Huntington 1957).

It has been long understood that militaries have their own institutional preferences, cultures, and standard operating procedures (SOPs) and that sometimes those preferences are in conflict with that of the leadership (Allison & Zelikow 1999). Whether the military or the civilian leadership controls the balance of political-military power in the state has real implications for foreign policy (Brooks 2008). However, this variable is not always stable and even in cases where full civilian control is assumed, the military will still have the tendency to follow their institutionally formulated SOPs rather than the 58 dictates of the leadership (Allison & Zelikow 1999). Even in the U.S. case, where civilian control over the military is widely assumed to be complete and constant, military officers, especially at the highest ranks, will not blindly go along with strategies the president proposes if they believe these strategies are not the most suitable for successfully completing their missions. Rather, the military leadership is likely to turn to the

Congress, the media, and other high-profile outlets to push their own agenda (Herspring

2005).

The intensity of the preference divergence between the civilian and military branches, regardless of cause, can play a key role not only in the domestic relationship between the leader and the military and in the distribution of goods within the polity, but also in the ability of both leader and the military to make accurate strategic assessments

(Brooks 2008) and it clearly inhibits the ability of the leader to achieve vital policy goals.

Preference divergence, at minimum, can cause the military to obstruct policy action on security objectives that the leader calculates is vital keep him in power. This divergence can also cause the leader to take action to prevent the military from carrying out SOPs or operations that are contrary to the preferences of the leader. At the extreme end of preference divergence, what this study is concerned with here, the internal conflict can influence the leader to channel resources into parallel military forces and intelligence services in order to achieve vital policy objectives, while it can lead the military to attempt to overthrow the leader.

The conventional wisdom sees terrorist groups as force multipliers and as a way to augment existing conventional forces where a unified state is seeking “security on the 59 cheap” and mobilizing conventional forces is costly. Indeed, even scholars who explain state sponsorship as a problem stemming primarily from domestic instability, leadership legitimacy, or as a situation where the leader is exporting the repressive tendencies of the state overlook the institution of the military as a key domestic player that is heavily invested in the outcome of internal struggles over governance, security policy, and resource allocation.

However, for many leaders, the biggest threat to their rule comes not from their external rivals, but from internal opposition (David 1991) in the form of the military

(coups d’état), the population (terrorism, insurgency, revolution, riots, mass demonstrations, voting for opposition parties), or from other elites (especially the ambitious within the leader’s current ruling coalition).35 While threats from the population can take various forms, their attempt to alter or overthrow the existing leader cannot be successful without the support (or absolute collapse) of the military. Even in the recent revolutions across the Arab world, where the military defected from the leader and supported the protestors (Tunisia and Egypt) the revolutions were successful at deposing the leaders, while in other states where the military stood relatively solidly behind the regime, the leader maintained power in the face of the revolt (Syria and

Libya)36 and the civilian death toll continued to climb.

35 Clearly elites and populations also matter, but I seek to develop a parsimonious first-cut of the intersection of domestic and international pressures on a leader’s propensity to sponsor terrorism. Given the greater ease with which the military can overthrow the government and the lack of attention to it as a variable in the literature on state sponsorship, the military will be the specific focus of this study. 36 Libya of course, is not a clear-cut case because of international intervention. It is not much of a stretch, however, to imagine a counter-factual situation where the international community did not intervene and it took a path similar to Syria. 60 Regardless of whether the military has direct or indirect influence in the internal politics of the state, “military leaders can influence the coalition and threaten the support base that keeps the political leader in power” (Brooks 2008, 31) and by extension, can also effect the chances that rival elites have at attaining and holding on to power. The military, then, is a key actor in both the internal and external conflicts of the state. That scholars have not examined the relationship between the military and leader in the context of state sponsored terrorism is baffling.

Omnibalancing and the Military The Theory of Leadership Survival posits that there are two ways that the military affects a leader’s decision to sponsor terrorist groups. First, when military leaders make up part of the winning coalition that keeps the leader in power (Bueno de Mesquita et al.

2003), then the military becomes a key faction the leader has to appease and/or balance in order to maintain that power. Appeasement may take many forms, however, in small coalition regimes, Bueno de Mesquita et al. (2003) notes that private goods are most likely to be distributed to elites in order to maintain loyalty to the ruler. In the case of the military, this can manifest as paying off the senior officers, providing high-tech military equipment to their respective forces, and providing high level jobs to relatives of the military brass. Indeed, there are indications that this appeasement tends to be quite effective in securing the loyalty of top military officials. However, private goods are limited and their distribution affects the war fighting capabilities of the military (Bueno de Mesquita et al. 2003). In general, this method of appeasement, due to limited resources, also cannot extend into the ranks of the junior officers, which are those most 61 likely to mount a coup against the regime (First 2012; Norden 1998; Zack-Williams &

Riley 1993).

Many of these leaders not only face the threat of a coup d’état, a clear indication of extreme preference diversion and internal threat to the regime (Desch 1999), but also have come to power themselves through coups, revolution, or other types of adverse regime change. Thus, they are acutely aware that additional balancing measures must be undertaken in order to secure their regime. Quinlivan (1999) argued that especially among the coup-prone regimes of Iraq and Syria, a system was devised to counter- balance the military that essentially made the leaders “coup-proof.” He further argued that leaders who came to power via coup are more likely to institute five structural elements of coup proofing so as to ensure their survival and make them less susceptible to future coups attempts. First was the widespread use of family, ethnic, and clan or religious ties and loyalties to tie subordinates personally to the survival of the regime; second was “the creation of parallel militaries that counterbalance the regular military forces;”37 third, the leader created a system of overlapping security agencies that “watch everyone, including other security agencies;” fourth, they encouraged professionalization and “expertness in the regular military” with intention to orient them to the task of national defense; and finally the leader made it clear that he prioritized the funding for all these programs (Quinlivan 1999, 135). All of these measures have intention of cultivating a conventional military that is moderately capable of external defense while protecting

37 I argue here that terrorist groups function essentially as one of these parallel militaries that work to counterbalance the traditional military and empower the intelligence agencies at its expense. 62 the regime from one that is strong, organized, and insulated enough to launch a coup d’état.

The method of a leader’s power attainment has implications for her mandate, legitimacy, and policy direction. While even a democratically elected leader is under immense pressure not to replicate the policy failures of the previous regime, those who come to power via coup d’état face a very real possibility of losing power the same way.

Furthermore, leaders who come to power as a result of a coup (revolution from above) or revolutionary struggle (from below) are much more likely to have aggressive foreign policies that are designed to legitimize the regime, export the revolution, and tend to be overly concerned with the potential that other states will act against it to overturn the revolution (Walt 1996).38 Furthermore, leaders in these cases often have to demonstrate to the remaining coalition that they are the most hardline of the regime and thus the most legitimate person to rule (Walt 1996).39

Appeasing the upper echelons of the military therefore is a necessary, but not sufficient, step for a coup prone leader to remain in power and diffuse the threat of coup.

Alternative military and security forces must also be cultivated that can do some of the job of the official state military without forcing the leader to rely on the military itself.

Terrorist groups, and the intelligence services that arm and equip them, provide the leader

38 In part, Walt argues, this is because adverse regime change alters alliance patterns as states sever or create diplomatic ties and the information available to all leaders about the capabilities and intentions of all states are significantly lessened. It also creates a winning coalition that is “revolutionary,” pushing leaders to demonstrate a revolutionary credibility in their domestic and foreign policies. 39 Sponsorship provides a way to signal this “hardline” to the remaining coalition. 63 with a way to circumvent the military leadership to attain vital policy objectives and secure their place in power.

The second way that the military affects a leader’s decision to sponsor terrorist groups is when these internal pressures from the military coincide with external pressures in a heightened external threat environment. The response to an external threat primarily falls upon the shoulders of the state’s official conventional forces. When the military mobilizes for external defense, the additional resources—especially the rapid increases in defense spending—and even the combat experience itself tends to have the effect of strengthening the unity, capabilities, and domestic political power of the military (Desch

1999). This power shift increases the military’s ability to mount a successful coup d’état and intensifies the internal threat posed by the military to the leader.40

Both of these arguments draw from the notion of omnibalancing (David 1991), which helps to clarify and unify the various explanations given by state sponsorship scholars for why a leader would sponsor terrorist groups. Indeed, when a state is viewed as unitary then a certain set of motivations to adopt particular, traditional, strategies for balancing and achieving “security on the cheap” emerges. However, when the state is not seen as unitary, but rather the state leader is understood to face threats to his rule from both internal and external sources, then it becomes clear that he face incentives and constraints that differ from the ones posited for the unitary state. He must opt for the right strategy in order to stay in power given those circumstances, even if it is not in the

40 Even in situations where leaders do not face an extreme internal threat, the preference divergence over security policy and increased domestic political power of the military in these situations can threaten the potential of the leader to attain vital policy objectives and weaken the legitimacy of his rule. 64 strategic interest of the state. In this case, and contrary to conventional wisdom, terrorist groups are not a way for a leader to augment conventional militaries, but rather a way for leaders to employ alternative force structures (also in the form of the intelligence services that arm and equip them) that offset the threat from (balance) the military in the face of cross-pressures to their rule. As shown below, autocratic regimes are especially vulnerable to this convergence of threat.

REGIME TYPE, SIZE OF THE WINNING COALITION, AND SUPPORT FOR TERRORISM The observation that authoritarian regimes are more likely to support terrorist groups is prominent in the literature on state sponsorship. Regimes with weak legitimacy and fractured governance structures are also considered to be more likely to sponsor.

How exactly these domestic variables translate into support for terrorism, however, remains underspecified in the literature. The Theory of Leadership Survival helps to make sense of this observation by drawing, in part, from Buena de Mesquita et al.’s

(2003) selectorate theory, which explains that every leader is kept in power by the support of a winning coalition in exchange for a share of private goods. The size of this winning coalition varies from quite large in democratic regimes, to exceptionally small in hereditary monarchies. Military regimes also have small coalitions and has the added characteristic that there are few outside the military establishment that can “make or break” the incumbent leader. Regardless of the size of the ruling coalition, however, the military nearly always plays a key role as part of the coalition and in order for the leader to be able to achieve vital foreign policy goals, he and the military must either have a convergence of preferences related to security policy or he must employ methods to 65 appease the upper echelons of the military while outright balancing the institution as whole.41

The military is not the only important component of the winning coalition, although they are the key focus of this study, and any policy option clearly generates winners and losers within the regime and the society at large. The military leadership has to be appeased enough to stay within the coalition, i.e., “win” enough to not defect and stage a coup, and must be kept strong enough to face external threats. Yet, other powerful elements within the must believe that their interests are being served, so as to not undermine the foundation of the leader’s support. When it comes to security policy especially, a “win” for other portions of the leader’s coalition may spell “loss” for the military, and vice versa. While this study does not propose to develop a theory of the conditions under which certain preferences will prevail, it is clear that this competition within the coalition creates security problems for the leader, who must appease and balance the various factions--keeping a sharp eye on the ability of the military to overthrow the regime--all while attempting to pursue the policy objectives that are vital to maintaining her rule. Terrorist groups provide the leader with a unique opportunity to achieve these goals and secure her position of power. The effects of these strengthening and weakening trends vary significantly by regime type. Not because of any “moral” advantage by democracies, but merely because of the structural effects of the size of the winning coalition.

41 Even in rare cases where the military does not make up a part of the coalition, these approaches (preference convergence, appeasement, and/or balancing) must still be taken as the military forms a key pillar of a ruler’s ability to protect his or her interests abroad and protect the security of the state (and their own rule) from external threats. 66 The size of this coalition is an important determinant for regime behavior and especially, “the survival of leaders in small winning coalition systems depends on their ability to provide private goods to their supporters” (Buena de Mesquita et al. 2003, 102).

While in large coalition systems, these private goods are distributed so widely, they become essentially public goods and leader survival becomes predicated on policy success.42 Democracies tend towards being large coalition regimes with large winning coalitions and are less likely to be coup-prone, which helps explain some important variance between sponsorship behaviors among leaders as well as the marked association between autocracy and state sponsorship.43

TERRORIST GROUPS, INTELLIGENCE AGENCIES, AND LEADERSHIP SURVIVAL In situations, then, where a leader faces the combination of an internal threat from the military and an external threat from an adversary state, terrorist groups, and the intelligence agencies that arm and equip them, provide the leader with an opportunity to circumvent the conventional military and pursue vital policy objectives which works to minimize these threats. Terrorist groups fortify the leadership in three ways, first, they provide cheap, on the ground intelligence to the regime, second, they provide a cost effective means of power projection, third, they allow the leader an avenue to achieve vital external policy goals within a structure that is accountable directly to the leadership

42 Indicating that in democracies or large coalition regimes, the connection between the leader and state sponsorship is more likely to be explained by the need to in avoid policy failure rather than the threat of military coup. 43 In some cases, especially in these small coalition regimes, sponsorship can also work to secure the leader’s position vis-a-vis this winning coalition; when the elite preference is for sponsorship, it can be used as a domestic signaling device by the leader. 67 rather than the military providing a key mechanism through which intelligence agencies are bolstered as alternatives to the military. Each of these will be explored in turn below.

Intelligence While most examinations of state sponsorship focus on the intelligence that is given to terrorist groups by states, there is also a flip-side to this coin in that terrorist groups are on the ground—normally in states that are adversaries to the sponsors—and the groups are able to provide key information regarding troop movements, take the pulse of the street, provide information on diplomat locations and patterns, scope hard and soft targets, and otherwise be “eyes and ears” on the ground without the regime having to invest in the much more expensive human and/or signal intelligence. These groups, as in the case of Hezbollah, often work in cooperation with the sponsoring state’s embassy staff and intelligence agencies (Ranstorp 1997) both receiving and transmitting vital information that can assist the leader in achieving vital policy goals necessary to maintain power all without involving the infrastructure, consent, or cooperation of the conventional military.

Power projection In addition to being the “eyes and ears” on the ground, these groups are also generally equipped with any number and quality of military hardware—supplied almost exclusively by the state’s intelligence services.44 Given their positioning, which can easily be understood as forward deployment into enemy territory, terrorist groups can

44 In fact, in the course of my research, I came across only one leader, Fidel Castro, who used his conventional military to train and equip external terrorist groups. 68 strike at vulnerable targets, cause crises of legitimacy for the enemy’s regime, and even attack military bases, supply lines, and outside sources of support to the adversary. The investment of funds, training, weapons, and travel assistance to terrorist groups is still fundamentally cheaper than other methods of power projection and the intelligence agencies that work in close cooperation with the terrorist groups tend to have few restrictions on movement or operations.

Perhaps most importantly, all of this is accomplished without involving the conventional military of the sponsoring state. While most accounts of state sponsorship argue that the military is not used because of the threat of war or because it is weak, this theory argues that the issue is fundamentally one of intentionally keeping the military weak enough not to oppose the leadership (though strong enough to protect the state) while still being able to achieve the foreign policy goals that are vital to the survival of the leader’s regime. In other words, terrorist groups do not augment existing conventional militaries they operate in spite of and in opposition to them.

Accountability While many scholars have addressed the “proxy” relationship between terrorist groups and the state and some, such as Byman & Kreps (2010) have gone even further and have begun to tease out the exact nature of the relationship through a principal-agent lens, the “sponsor” remains the unitary state. If the state is disaggregated, however, the question becomes, who exactly is the principal? This study makes the assumption that this principal is the same actor who is ultimately responsible for foreign policy decisions, that is, the state leader. Terrorist groups by answering solely to the leader, or the 69 intelligence services that do, serve as an additional method through which the leader can signal to important members of his coalition that vital interests are being attended to. In addition, by being accountable only to the leader of the state45 or the intelligence services that are, the terrorist group bypasses the conventional military establishment. This adds another layer of protection for the leader as he struggles to balance against the convergence of internal and external threats, especially in the context of small coalition regimes.

Again, it is important to keep in mind that while there is an association between

(small coalition) autocracies and state sponsorship, it is vital to differentiate the regime type from the threat environment that the leader faces. While autocratic regimes are predisposed to face High External & High Internal (HEHI) and Low External & High

Internal (LEHI) threat environments, they are exceedingly more likely to sponsor terrorism in an HEHI than LEHI threat environment. Indeed, this study will demonstrate through both the quantitative statistical analysis in Chapter 4 and the qualitative case studies of Chapter 5 that it is not the regime type that is the key determinant for sponsorship but rather the convergence of internal and external threats.46

SPONSORSHIP IN DEMOCRACIES Existing literature on state sponsorship addresses state sponsorship by democratic regimes rarely, if at all. Yet, the data indicate that democratic leaders are also likely to

45 Again, this assumes the conditions of “onset” of state sponsorship. In some cases, most notably Pakistan and Iran, these intelligence services and the terrorist groups they support have maneuvered into a position of power that ultimately poses a threat to the leadership in a similar way that the military did previously. This shift and perhaps “maturity” of state sponsorship is an avenue for future research. 46 The correlation coefficient between the HEHI threat environment and autocracy was 0.2148. 70 support terrorism. Of the 967 leaders included in this study, 543 of them were coded as democratic47 and 73 of those sponsored one or more terrorist groups. While this occurrence is still quite low (13.44% of democratic leaders and only 7.54% of all leaders in the study) it is certainly not null.

The Theory of Leadership Survival specifically addresses situations where leaders face a combined internal and external threat to their regime. Yet, most democratic regimes do not experience the same level of coup threat, internal war, or failure of state authority that other regimes do.48 The basic assumptions that guide this argument still apply, but are diluted: 1) these are rational leaders with the desire to maintain power; 2) they face the situation where their legitimacy stems, in large part, from being able to overcome the policy failures of the previous regime; and 3) the military, due to preference divergence over security policy, poses a lesser threat than in cases of a coup d’état, but still function as an obstacle to the leader. Specifically, the military can prevent a leader from achieving policy success in an issue area that is of critical importance to her ability to maintain office. This study expects that democratic leaders that sponsor terrorist groups, despite the lessened threat of coup, do so in the context of an environment that is characterized by policy divergence with the military and extreme electoral (legacy) pressures that are being made worse by military obstruction.

47 While extensive literature exists on what exactly should be considered an “autocratic” or “democratic” regime, this study was primarily concerned with exploring autocracy as a causal variable and less so with defining clearly which regimes qualified as “democratic” and by which criteria. Therefore, these categories are very rough and break at the “0” on the Polity IV scale. That is, all leaders and cases considered autocratic scored a negative (-10 through 0) score on the scale. Democratic leaders were coded as having Polity IV scores of 1 through 10. 48 The data also indicates that democracies are over-represented in High External & Low Internal (HELI) sponsorship cases, explained in the HELI section below. 71 While the quantitative portion of this particular study is not “fine-tuned” enough to capture the policy divergence between the leader and the military at this sub-coup level of divergence, future works will take this further.49 Again, it is important to note that it is not the regime type that is responsible for state sponsorship in these cases, but rather the regime structure predisposes the leader to face particular kinds of pressure from the military in the form of variable preference divergence, that when combined with a heightened threat environment, especially at the extremes, incentivizes the leader to sponsor terrorism.

An examination of sponsorship patterns in democracies by threat environment show that it is indeed the threat environment that is the determinant, not the regime type.

Given that democracies are more likely to experience opposition from the military as policy divergence rather than outright coup, the over-representation of democratic sponsors in the HELI (High External & Low Internal) and LELI (Low External & Low

Internal) categories in Figure 3.1 below is not surprising.

49 For instance, while extreme policy divergence is captured here as “threat of coup” and signals a shift in the threat environment the leader faces, the distinction between policy convergence, minor policy divergence, or even substantial policy divergence between the military and leader are not addressed in the quantitative portion of this dissertation, though it is mentioned in examples and the case studies. It is likely, however that the High External & Low Internal sponsors experience substantial policy divergence between the military and leadership and this should be examined at length in future studies.

72

Figure 3.1: Number of Leaders and Sponsors in Democratic Regimes (1968-2001).

Proportionally, however, there is a tremendous difference, i.e., even in democratic regimes, HEHI (High External & High Internal) threat facing leaders were much more likely to be sponsors (31.4% of HEHI facing democratic leaders sponsored terrorist groups) than HELI (20.6%) or LELI (2.6%) ones. Again, while this study is not designed to test the nuances of policy divergence in the predominately democratic HELI threat environment, the logic of the theory remains the same.

SPONSORSHIP IN AUTOCRATIC REGIMES For many autocratic leaders, the threat posed by the military is not a matter of mere obstruction of policy because of preference divergence, but can manifest in direct

73 challenge to the regime through coup d’état.50 By its very nature, the structure of a small winning coalition makes the military a larger threat and requires that private goods

(spoils) be channeled to the military leadership in order to keep them in support of the regime (Bueno de Mesquita et al. 2003). Other factions within the regime have to be taken into account by the leader as well, however, especially those on which the leader depended to get into power in the first place. In the instance of a military regime, the faction that brought the leader to power is the military itself, but even then, a split tends to occur as the leader focuses on the interests and demands of the office, leaving the leader to mediate the factions within the coalition in a similar fashion as other small coalition regimes. Even when the military is not the key constituent responsible for helping the leader attain power, the structural nature (small winning coalition) of an autocratic regime, makes them a bigger threat to the leadership than they would be in a democratic one.

Again, however, the regime type merely predisposes the leader to face a particular type of threat from the military, albeit in this case it is the one that the Theory of

Leadership Survival is designed to handle best. In autocratic regimes especially, leaders tend to face a heightened threat of coup d’état, which, when combined with an external threat, creates the HEHI threat environment. While again, threat environment and regime type are conflated in the literature, autocratic regimes are most likely to both face HEHI

50 Autocratic leaders that sponsor terrorist groups are disproportionately represented in the High External & High Internal threat environment, as will be discussed below. 74 threats (80% of HEHI facing leaders are autocratic) and to sponsor terrorism (see Figure

3.2 below).

Figure 3.2: Percent of Autocratic Leaders that supported terrorism (1968-2001).

Note: Recall that sponsorship among Democracies was 13%. The correlation coefficient between Autocracy and Sponsorship was 0.1860.

It is an error, however, to equate regime type with sponsorship as both democracies and autocratic regimes provide support for terrorist groups. The section on the threat environments below will demonstrate that while regime type may predispose leaders to face particular threat environments, it is the threat environment itself that is the primary determinant of state sponsorship and when that threat environment changes, so does state support for terrorism.

75 THREAT ENVIRONMENT AND STATE SPONSORSHIP The notion that internal and external threat environments interact to shape relations between leaders and their militaries has been examined by Desch (1999) who looks specifically at these same variables, the internal security environment and external security environment, to structurally determine level of civilian control over the military.

While he has clear predictions for situations of High External51 & Low Internal (HELI) and Low External & High Internal (LEHI) threat environments, he concedes that High

Internal & High External (HEHI) threat environment, the key category we are examining here, is an “indeterminate” threat environment for his theory.

Figure 3.3: A Breakdown of the Threat Environment Leaders Faced (1968-2001)

51 Defined as the presence of at least one Militarized International Dispute during the observation year in the main dataset and as the presence of at least one during the tenure of the leader in the sub-dataset that was condensed to the leader level. This, and the internal threat specifications will be discussed more in depth in the following, methodological, chapter. 76 If the threat environment categories developed by Desch (1999) to examine the relationship between the civilian leadership and the military (Table 3.1) are used not to examine civilian control of the military, but rather a different aspect of civil-military relations, that is, when leaders fear or find their conventional militaries unreliable and instead empower alternative parts of the regime (especially their security services) to circumvent the military through the use of terrorist groups to achieve policy goals, we can see that a high external threat predicts a sponsorship in general, but when combined with an internal threat, this sponsorship becomes a vital component to regime survival.

The Theory of Leadership Survival predicts that the conditions of a High External

& High Internal (HEHI) threat environment create a situation that is most likely to compel the leader towards certain self-protective actions, including coup-proofing measures, and the state sponsorship of terrorist groups.

77 Internal HIGH LOW

HEHI: Sponsorship extremely likely. HELI: Sponsorship somewhat likely. Omni-embattled leader, security External threat, military circumvented service proliferation, military due to divergent security policy, HIGH circumvented due to high coup threat. military obstructs leader’s attempt to

achieve policy items vital to survival of regime.

External LEHI: Sponsorship unlikely. LELI: Sponsorship extremely unlikely. Internal threat takes precedence and Leaders face no discernable threat from

external terrorist groups cannot assist rivals or the military, if sponsorship LOW in relieving the pressure. Internal occurs it’s outside the bounds of the

“death squads” may be employed. theory.

Table 3.1: Theory predictions for sponsorship in each threat environment

These four threat environments will be examined in turn in the sections below.

Specifically, the predictions of the Theory of Leadership Survival will be examined in depth and the descriptive data on the various threat environments and the sponsorship frequencies of the leaders presented.

HIGH EXTERNAL & HIGH INTERNAL (HEHI) THREAT ENVIRONMENT In this threat environment, leaders face both an external threat from an adversary and the state has a very recent history of coup d’état52 from the military, indicating the leader is in a situation that calls for omnibalancing. In some cases, the leader manufactures an external threat, in part because “war and international conflict can

52 Attempted coups, plots of coups, and successful coups are all considered in this “history of coup d’etat.” 78 provide leaders unique opportunities to deal with their opposition” (Chiozza & Goemans

2004, 606).53 However, this tactic can be a double-edged sword. Many leaders, such as

Saddam Hussein,54 embark on an external adventure with hopes for easy victory and quick consolidation of domestic power, only to have it morph into a much larger conflict requiring the diversion of resources to build-up and mobilize their armed forces.

Here, then, leaders face a dilemma, regardless of whether they have instigated the conflict, they have to demonstrate to the rest of their ruling coalition that they are strong, capable, and are able to meet the external threat. Yet if they build-up, equip and mobilize their armies to meet this external threat, they also significantly strengthen these armies and heighten the already existing threat of coup d’état against their rule. It is in this context, especially, that we expect to see the leader build up parallel militaries, security services, and use terrorist groups as a way to respond to the external threat while only minimally expanding the institutional power of the military.55 The theory predicts that leaders facing this threat environment are most likely to sponsor terrorist groups.

There were a total of 177 leaders in the study facing a HEHI threat environment, including some of the most well-known state sponsors of terrorism in the world, such as

53 Indeed, while some recent studies have highlighted the fact that some personalistic regimes may experience higher instances of MIDs because they instigate them (Peceny, Beer, and Sanchez-Terry 2002, Peceny and Butler 2004) this would not change the basic conventional military mobilization that would be required. Whether or not the leader instigates the conflict may have implications for the relationship between a diversionary war effect of state sponsorship versus mobilizing in an unwanted crisis, yet that is a topic for further research and is outside the bounds of this current study. In short, the instigator of the MID should not affect the predictions of this model. 54 One of the confirming cases that will be explored in Chapter 5. 55 Where there is “no choice” but to expand the military, we expect that the leader will do everything possible to ensure the loyalty of the officer corps and will balance with parallel military structures as much as possible. 79 Saddam Hussein, Hafez al ‘Assad, Ayatollah Khomeini, and Mummar Qaddafi, most of which will be examined in depth in Chapter 5. Figure 3.4 below indicates that out of these 177 HEHI facing leaders, 66 of them (or 37.29%) sponsored terrorist groups, more than any other threat category.

Percent of Leaders Sponsoring by Threat Environement (1968-2001)

40.00%

30.00%

20.00%

10.00%

0.00% HEHI HELI LEHI LELI

Figure 3.4: Sponsoring leaders as a total percentage of threat environment facing leaders (1968-2001).

HEHI sponsors of terrorism are expected to have higher military expenditures than other leaders, including non-sponsoring HEHI leaders because of the need to mobilize internal resources to meet their external challenges as well as to appease the senior officer corp.56 As conventional armies mobilize and leaders are forced to divert more of the GDP to the military, it strengthens the military’s capabilities to fight an

56 This appeasement is done through high salaries, high-tech equipment that increases individual officer’s prestige, and other private goods. 80 external war, yet also increases the organizational and operational capacity for the military to overthrow the regime. Thus, increased military spending in HEHI, indeed all, threat environments is expected to predict higher threat instances of sponsorship.57

If HEHI threat facing leaders are most likely to sponsor terrorism and autocratic leaders are structurally predisposed to face more HEHI threats, this relationship may help explain the over-representation of autocratic regimes as state sponsors. HEHI leaders are also expected to be the most entrenched sponsors, given that their regime survival becomes bound up with this support.58 However, the figure below shows that even leaders in democracies are more likely to sponsor terrorist groups in HEHI environments.

57 This prediction is in stark contrast to San Acka’s (2009) expectation that state sponsorship arises from the inability to mobilize resources (internally balance) or attain external allies (externally balance) to meet an external threat. 58 This claim is in opposition to Navin Bapat, whose unpublished work on state sponsorship of insurgencies and internal politics suggests that while autocratic sponsorship should be more common, democratic sponsorship is expected to be more extensive. 81

Figure 3.5: Percent of Democratic Leaders that are State Sponsors, broken down by Threat Environment (1969-2001).

A full 34.29% of HEHI facing democratic leaders are sponsors of terrorist groups.

This is most stark in comparison to the 20.65% of sponsoring democratic HELI facing leaders, where democratic leader are much more prevalent. Thus, even across regime- type, the threat environment of the leader is a bigger predictor of state sponsorship.

HIGH EXTERNAL & LOW INTERNAL (HELI) THREAT ENVIRONMENT For this threat category, Desch (1999) is able to make some clear predictions for the overall relationship between the leader and the military. He argues that it will be one of relatively uncontested civilian control where the leaders are experienced and capable in security matters, the military is unified, externally oriented, and generally takes positions on security policy that are convergent with the leader’s. However, this study

82 argues that when the policy does diverge and the pressure on the leader to take action are acute, especially in cases where predecessors have lost power to the incumbent over these same policies, incentives can emerge to push for success on the policy issue at all costs.

For leaders of HELI threat environments, this policy divergence and high external pressure can incentivize support for terrorist groups.59

While a deeper exploration into the dynamics between HELI facing leaders, their militaries, and terrorist groups will need to be undertaken in future studies, one clear picture emerges from the figure below, that is, regime type is not, by itself, responsible for the state sponsorship of terrorism. Indeed, the difference between sponsoring autocrats and sponsoring democratic leaders was just over a single percentage point.

59 Despite the prevalence of this threat environment and sponsorship, this study is not well designed to test the propositions of HELI sponsorship. In future work, attention will be paid specifically to HELI sponsorship and the more nuanced relationships between the leadership and the military. 83

Figure 3.6: Sponsoring Leaders Facing the HELI Threat Environment broken down by Regime Type (1968-2001), regime type averaged over tenure.

Even when leaders are fairly secure from the threat of a coup d’état (Low Internal threat), when they experience high external threat, especially one that has the legacy of having caused the downfall of previous leaders and thus threatens their ability to maintain power domestically they will try to mobilize militarily to meet this threat. If the leader experiences preference divergence with the military over security policy—with the military having the ability to act as a veto player, a leader is likely to make the choice to circumvent the military’s foot dragging by shifting resources to other agencies (especially intelligence) and support the terrorist groups that can help him achieve the goals necessary to demonstrate competency to his coalition and remain in power.

While a closer look into the relationship between HELI threat environments and state sponsorship will largely be left for future research, there are readily available

84 anecdotal examples to help illustrate the dynamics. For example, in the early 1980s after the US pulled peace keeping troops out of Lebanon following the attacks on the US embassy and Marine barracks, the US Secretary of Defense Casper Weinberger had a clear aversion to deploying military forces in the name of Reagan’s policy of “rolling back communism” upon which Reagan had staked his credibility.

Reagan’s subsequent decision to circumvent Weinberger’s authority and support the Contras in Nicaragua and UNITA in Angola using the National Security Agency was undertaken in the interest of securing these policy issues that were vital to securing his legacy, legitimacy, and occurred in the context of a severe diversion of security policy between his administration and the military establishment. While the key elements of external threat (from the USSR and “international communism”) and internal threat

(albeit in the form of policy divergence with the military) to his legacy and legitimacy are present in this case, the Theory of Leadership Survival, as specified in this first cut, has a difficult time addressing these cases empirically. However, the basic logic still applies.

LOW EXTERNAL & HIGH INTERNAL (LEHI) THREAT ENVIRONMENT A third threat environment leaders potentially face is one with a low external threat from rival states and a high internal threat from the military. In this situation, according to Desch (1999), the civilian leaders are inexperienced and have only subjective control over a unified military that has an inward orientation. Furthermore, he argues that it is likely the civilian and military leadership have significant divergent ideas over security policy. However, because there is no external threat, even when additional resources are diverted to the military to help with conditions such as internal war, which 85 also increases the threat posed by the military to the leadership, there is nothing gained by enlisting the support of external terrorist groups.

The Theory of Leadership Survival predicts that few if any sponsoring leaders will emerge from this threat environment. Leaders will have few incentives to operate strategically, or otherwise, beyond their borders and despite having little control over their military forces, they will not view empowering intelligence services and employing external terrorist groups as useful for solving their particular security problems. Instead, leaders may enlist the help of domestic paramilitary groups or “proxy militias” within the state. Importantly, new research shows that these domestic paramilitary groups also function as a way for the leader to avoid giving power to a centralized military that is likely to perpetrate a military coup (Ahram 2011).

The Theory of Leadership Survival predicts that given these constraints and incentives, the LEHI threat environment will produce little if any instances of state sponsorship and that it will not be a significant predictor of a leader’s choice to provide support to terrorists.

86

Figure 3.7: Leaders and sponsors in the LEHI threat environment (1968-2001)

Indeed, the data indicate that only 6% (that is nine out of the 148 LEHI facing leaders) were sponsors of terrorism and only .09% of total leaders were LEHI facing sponsors. While leaders facing this threat environment are expected to be overwhelmingly concerned with the internal threat to their rule, a situation similar to that of HEHI facing leaders, in this case there is no pressure on the leader to omnibalance and thus, providing support to an external terrorist group cannot help alleviate the threat the leader faces.

LOW EXTERNAL & LOW INTERNAL (LELI) THREAT ENVIRONMENT The final threat environment leader face is one of both a low external threat from rivals and low internal threat from their militaries. In this situation, the Theory of 87 Leadership Survival expects that in the absence of these cross-pressures leaders will experience no incentive to sponsor terrorist groups. The LELI threat environment is not expected to be a significant predictor of state sponsorship.

Figure 3.8: LELI facing leaders and percent that are sponsors (1968-2001)60

Indeed, the data indicate that only ten of the 289 LELI facing leaders

(approximately 1% of all leaders) sponsored terrorism. The ten instances of sponsorship that did emerge from this category will be examined as outliers in Chapter 6. It will be shown that the bulk of these ten leaders either served less than one year in power and/or had predecessors that face a different threat environment. The predominant outlier is Tun

60 To be coded as an LELI sponsor, they could not have experienced any internal (coup) or external (MID) threat over the course of their rule. 88 Abdul Razak’s Malaysia and his long-time provision of support to the Moro National

Liberation Front.61

The understanding of state sponsorship that emerges from reorienting the focus to the state leader, adding the military as a key factor, and examining the threat environments that leader faces and must respond to allows a number of testable hypotheses to be developed, a task undertaken by the next section.

HYPOTHESES The insights from the state sponsorship and civil-military relations literatures, along with advances in foreign policy theory, inform the Theory of Leadership Survival developed here. While taking into account the well-established pattern of leaders adopting sponsorship as a way to deal with external crisis, this theory adds the dimension of a simultaneous internal threat to their leadership, and thus the need for the leader to omnibalance in order to retain power, that can establish an empirical foundation for understanding state sponsorship and can shed light on some of the most notorious cases.

While high external threats, in general, put unique pressures on leadership survival and in cases of policy divergence with the military are likely to lead to support for terrorism, the

Theory of Leadership Survival predicts specifically that:

Hypothesis 1: Leaders facing High External and High Internal threats simultaneously (HEHI) are most likely to sponsor terrorism.

61 This case will be examined in depth in Chapter 6. 89 In this case, external threat refers to Militarized International Disputes (MID’s) as will be discussed in the next chapter, while internal threats are explicitly understood to be threat of coup d’état. The idea that an “internal threat” (left undefined) leads to state sponsorship has been examined by scholars such as Byman (2005) and Arnold and

Livingstone (1989) who see it as a threat to the legitimacy of the regime. While this study focuses primarily on the threat of military coup as the primary internal threat,62 as outlined by Desch (1999), there is a question as to whether the destabilizing and internally mobilizing effects of internal war should also be considered. Thus, an internal war dummy will be used to control for the possibility that it, rather than coup d’état is the internal threat leading the leader to sponsor terrorism. While it is likely to add to the strain a leader faces in preserving her rule and is thought to account for some of this policy decision, internal war is not expected to be a primary determinant of sponsorship.

62 Given the lack of quantitative data available on intelligence agencies’ budgets and operations, this study does not attempt to make claims regarding the actions of these agencies. While the theory has been created inductively and the intelligence agencies will play major roles in the exploration of the case-studies in Chapter 5, the focus for the quantitative portion of the study (i.e. hypothesis testing) will remain on the threat environment faced by the leader (specifically the external threat by state adversaries and internal threat by the military), regime type, and military expenditure. 90 Leaders facing Internal War and State Sponsorship (1968-2001)

32% Sponsors Non-Sponsors 68%

Figure 3.9 Leaders facing Internal War and their relationship with State Sponsorship 1968-2001.

While it is likely that regime type merely predisposes a leader toward facing a particular threat environment, the small winning coalition structure of autocratic regimes and nature of the role that the military can play in these circumstances lead to the second hypothesis:

Hypothesis 2: Autocratic leaders are most likely to sponsor terrorist groups.

Finally, and contrary to the existing literature that argues leaders sponsor terrorism because they have weak, and underfunded, conventional militaries, this theory expects that increased military expenditure is required to meet the external threat, but also

91 further equips the military to operate as a veto player and to pose the threat of internal coup and that in these circumstances, sponsorship will increase in tandem with increased military expenditure.

Hypothesis 3: Increased military spending leads to an increased likelihood of state sponsorship.

This explanation of conditions provided by the Theory of Leadership Survival provides leverage to many of the conventionally accepted, and oddly non-conflicting, explanations as to why states sponsor terrorism. Importantly, by adding in the notion of omnibalancing and the fact that leaders experience and respond to threats in a different way than a unitary state is expected to, it can also account for the puzzle of why states sometimes continue to sponsor terrorist groups well beyond when it appears to be in their interest to do so (Byman 2005). Indeed, many leaders continue, and sometimes increase, their support for terrorist groups even after they have been subject to massive counterterrorism efforts, including military attack (Qaddafi in Libya), multilateral sanctions (Saddam Hussein in Iraq), and isolation from the international community

(Khomeini in Iran and Qaddafi in Libya).

Together, these three hypotheses and alternative explanation of responding to internal war will be tested in Chapter 4 in an attempt to develop an empirical foundation for the study and understanding of the state sponsorship of terrorism.

92 CONCLUSION The explanations commonly given for why states sponsor terrorism can be understood as corresponding closely with the international and domestic levels of analysis. These explanations are not competing, but rather can be combined into a clear and operationalizable framework for understanding the decisions of state leaders to engage in state sponsorship. When superimposing the state sponsorship literature on that of foreign policy formulation and civil-military relations, we see clear parallels and can begin to understand how the pressures of internal threats to the regime from the military and external threats from international conflict combine to induce the leader to sponsor terrorist groups to fortify themselves in power.

The next chapter takes a closer look at the hypotheses generated by the Theory of

Leadership Survival. First, it explains why the existing data was not appropriate for the examination of state sponsorship and explains the rationale for the creation of an original dataset for this study. Second, it will show how the variables have been operationalized and explain the methods and criteria for the data collection for the new dataset on state sponsorship. In addition, it lays out the methods that will be used to test the hypotheses using a large-N analysis. Finally, it presents the results of the descriptive and inferential analysis to examine how well the data adheres to the theory outlined in this section. The fifth chapter will address more closely the causal mechanism uncovered here and while the sixth will allow an examination of outliers in a case comparison of threat environments and sponsorship choice.

93 Chapter 4: Empirical Support for the Theory of Leadership Survival

The results of the cross-section time-series analyses conducted and presented in this chapter and the multitude of descriptive data presented throughout the study indicate that there is empirical support for the hypotheses that the HEHI threat environment is a significant predictor of state sponsorship. The analysis also indicates that increased military expenditure (as a % of GDP) is also a strong predictor of state sponsorship, even when controlled for the tremendous military spending of the superpowers. Autocratic regimes were also found to be more likely to sponsor terrorist groups. The data, methods, and detailed results will be discussed at length below.

The chapter is organized as follows: first, it explains why the existing data was not appropriate for the examination of state sponsorship and explains the rationale for the creation of an original dataset for this study. Second, it will show how the variables have been operationalized and explain the methods and criteria for the data collection for the new dataset on state sponsorship. In addition, it lays out the methods that will be used to test the hypotheses using a large-N analysis. Finally, it presents the results of the descriptive and inferential analysis to examine how well the data adheres to the theory outlined in Chapter 3.

LACK OF EXISTING EMPIRICAL FOUNDATION One of the major obstacles to formulating effective policy to counter state sponsored terrorism is the lack of empirical data and analytical work regarding its causes, characteristics, and broader circumstances. This has led to real-time experiments on the effects of sanctions; export controls, covert action, military strikes, and even regime 94 change. These policies have been enacted with little theoretical or empirical underpinning and have aimed primarily to drive up the cost of sponsorship to the levels where the benefits were no longer attractive. However, we have only anecdotal and speculative information regarding whether or how these levels of acceptable cost or benefit vary. As the incentives and constraints that leaders face when making the decision to begin sponsorship become more empirically clear, counterterrorism policies can be more effectively designed and implemented and potential future sponsors provided with alternatives.

No comprehensive dataset on state sponsored terrorism exists for use by scholars and policy makers. Although the U.S. Department of State, the National Counterterrorism

Center (NCTC),63 and the RAND64 Corporation maintain databases on terrorist attack incidents and the U.S. State Department publishes an annual Country Reports on

Terrorism they do not provide comprehensive information or usable datasets. Meanwhile, the most commonly used datasets in the field, ITERATE65 and the National Consortium for the Study of Terrorism and Responses to Terrorism’s (START) Global Terrorism

63 NCTC maintains a Worldwide Incidents Tracking System, which catalogs characteristics of the attacks themselves, but has little to no data on the perpetrators. Available at http://www.nctc.gov/wits/witsnextgen.html 64 RAND Worldwide Terrorism Incidents dataset no longer considers state sponsored terrorism a terrorist attack. In previous editions, the default setting was ‘no’; however, if a state was behind/responsible for an attack and this claim was deemed to be credible then ‘yes’ was entered. http://www.rand.org/nsrd/projects/terrorism-incidents/about/definitions.html accessed September 3, 2011. 65 Mickolus, Edward F. INTERNATIONAL TERRORISM: ATTRIBUTES OF TERRORIST EVENTS, 1968-1977 [ITERATE 2] [Computer file]. ICPSR07947-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [producer and distributor], 1982. doi:10.3886/ICPSR07947. 95 Database (GTD),66 either have an underdeveloped state sponsorship variable that is not useful for analysis, or lack the variable completely. Two more recent datasets, the

Nonstate Armed Groups dataset developed by Belgin San Akca and the Non state Actors and Alliance behavior dataset developed by Navin Bapat do include state sponsorship as a variable, but they are not well suited, as will be explored below, to answer the question posed by this study.

The next section examines each of these datasets in depth, explains why they are inadequate to address the research question of under what conditions will leaders provide terrorist groups access to state resources. The remainder of the chapter will then be dedicated to explaining the research design, construction of the dataset developed for this study, the operationalization of the variables, the method of analysis, and the results of the regressions and other inferential analyses.

EXISTING DATASETS There are a number of existing open source datasets available for researchers who study terrorism, although four in particular are were particularly relevant for this study: the ITERATE series of datasets on terrorist events, the START database on terrorist groups and the GTD data on terrorist attacks, San Akca’s Non State Armed Groups dataset, and the Non state Actors and Alliance behavior dataset by Navin Bapat. None of them are appropriate for examining the phenomenon of state sponsored terrorism, especially from the perspective of the leader, as the section below will illustrate.

66 The Global Terrorism Database (GTD) is one of the most comprehensive and reliable archives of attack incident, perpetrator (including suspected), target, and causalities. Available at http://www.start.umd.edu/gtd/ 96 ITERATE (1-4) The International Terrorism: Attributes of Terrorist Events (ITERATE) datasets are the most widely used open-source attack incident dataset in the terrorism literature.

The ITERATE common file datasets do code for state support in attacks when the state either provided “direct support to terrorists in incident” or “officially-sanctioned participation of government personnel in incident;” however, all other forms of state support are coded as “irrelevant” (Mickolus et al 2004, 17). The variable of state support, as coded by ITERATE, is not adequate for this study for two reasons. First, the participation by government officials in terrorist attacks fall under the banner of state terrorism67 and not state sponsorship. Second, while some states do provide direct support for particular incidents, the active provision of arms, weapons, training, passports, logistical and organizational assistance, safe haven to terrorist groups outside the context of a particular incident is equally vital to consider and is excluded from the ITERATE data. In fact, Overgaard argues that terrorist groups gain “direct utility from the peaceful use of its resources” (1994, 454) indicating that even support provided by states that does not go directly to supporting a particular terrorist attack will still benefit the group and increase their overall capabilities to perpetrate future attacks.

Thus, the only existing event dataset on terrorism that explicitly includes state sponsorship as a variable both incorporates aspects of state terrorism that should not be

67 State terrorism was specifically excluded from the analysis in Chapter 2, Figure 2.1 97 included,68 while under-representing cases of state sponsorship by restricting the definition to provision of support only within the context of a particular incident.

START and the GTD The National Consortium for the Study of Terrorism and Responses to

Terrorism’s (START) Global Terrorism Database (GTD1: 1970-1997 & GTD2: 1998-

2004) was developed in an attempt to code and verify existing open source information on terrorist attacks. While the GTD specifically indicates that they do not include state support as a variable, START also maintains a Terrorist Organization Profile (TOP) database which, in addition to listing the relationships that the groups have with other groups, also includes some information on whether the group has been provided with support by a state, which state provided the support, and in many instances, which type of support and during which rough time period the support was provided (Miller & Smarick

2012).

Thus, although this data is incomplete—and focused on the incident and terrorist group, not the state-- it provides a good initial point of cross-reference for this project.

However, this research design mandates more specific information than to know that

Libya “supported” a certain rejectionist group during the 1980s or that Iran provided money to the Palestinian Islamic Jihad. In order to get a clear picture of the relationship between state support and the threat environment of the leader, we need to be able to pin point which years the state provided support, what type of support was offered, and to which groups.

68 Given the criteria outlined for this study. 98 Non State Armed Groups (NAGS), their Targets, and Sponsors The third database considered for this study, Belgin San Akca’s Non-state Armed

Groups, Their Targets, and Sponsors (NAGS), will receive the most attention since it is the closest potential contender for this study and also seeks to correct two most salient problems in existing datasets that address state sponsorship. First, those that do include state sponsorship include it only as a minor variable, and second, these datasets were all developed to speak to theoretical questions other than state sponsorship (San Akca 2009).

Using the target-group-potential sponsor triad as the unit of analysis, she utilized the UCDP/PRIO Armed Conflict dataset to determine the group and target state and then used the target state’s politically relevant group to determine potential supporters (Maoz

1996). Using the years the group was active as a time frame for the triad and open source data sources, she then collected data on whether the member states of the target state’s politically relevant group were providing support to the non-state armed group.

There are three major research design problems in this dataset that I have avoided by using the leader-year as the unit of analysis. First, the politically relevant group of a target state, which is used to determine the potential supporter, consists only of the states contiguous to the target state, regional powers, and global powers which have the ability to project power outside their regions. Given these constraints the dataset cannot, and does not, examine the relationship between triads outside these conditions,69 which excludes many key anecdotal cases. Thus, while Cuban support to the MLPA in Angola, and Saudi contributions al-Qaeda in Taliban-controlled Afghanistan would not be

69 For a discussion of disadvantages of using politically relevant groups see Moaz & Russett 1993; Maoz 1996; and Lemke and Reed 2001. 99 included in the dataset, Libyan support to the IRA in the UK would be because, as a

Great Power, the UK is part of Libya’s PRG (San Akca 2009, 50). The focus on PRGs also overlooks one of the major reasons states support terrorism in the first place: terrorist groups provide states the ability to project power far beyond their borders. By using all states in the international system as the potential sponsor, by leader-year, this study is able to include all of the cases San Akca accounts for in her triads as well as these vital cases she omits.

Second, by only including the years that the group was active in the dataset, it is unable to account for the changes in the internal and external situations of the state that may have led to the onset of support. This also makes any analysis of the change or stability in sponsorship patterns between the Cold War and post-Cold War period difficult. Finally, the some states, such as Iran, Libya, Syria are known to support multiple groups against the same target states (i.e. Israel or the US), these cases, and their attending circumstances, will be over-represented in San Akca’s dataset. This creates a serious potential for bias in the analysis and in the interpretation of the results. By placing the focus on the incentives and constraints of all states (specifically their leaders), as potential sponsors, within a specified time period and coding the type of support provided, rather than a state’s relationship with any particular terrorist group, it will allow a clearer picture of the phenomenon.

Non-State Actors and Alliance Behavior Dataset The final dataset considered for use in this study was the Non-State Actors and

Alliance Behavior (NSAAB) dataset developed by Navin Bapat (2007). Like the dataset 100 created by San Akca, Bapat began with the UCDP/PRIO Armed Conflict dataset to determine the group and target state conflict and then created a list of potential supporter/group dyads using Rasler & Thompson’s strategic rivalries data (2006). The dataset assumed that potential supporters were those most likely to be in a rivalry with the target state and thus to use the group as a tool within this rivalry.

Using the rivalry-year as the unit of analysis, Bapat’s coders examined Keesing's

Record of World Events, Lexis-Nexis, Facts on File, and the SIPRI Yearbook for the relationship between the identified group and the target state’s rivals in cases identified by Byman et al. (2001). The coder identified active and intentional support from the rival to the militant group the variable was coded as 1, and 0 otherwise (Bapat 2009, 21).

There were several attributes to the dataset, which makes it inadequate for the purposes of this study. First, the dataset only covers 1989-2001, thus excluding nearly 20 years of known state sponsorship cases70 and making any comparison of the incentives and constraints faced by decision making leaders during the Cold War and post-Cold War eras impossible. Second, rather than assuming that a target state’s rivals are more likely to sponsor terrorist groups, and selecting cases on this assumption, this study maintains that rivalry and conflict are variables that, while theoretically related to state sponsorship, have yet to be adequately empirically tested and do not make for unbiased case selection criteria.

Thus, there are no open-source datasets available to scholars that are properly designed to address the specific question being posed by this study, under what

70 Including the key cases of “classic” sponsorship that occurred during the Cold War. 101 circumstances will state leaders make the decision to provide state controlled resources to terrorist groups. As Crenshaw notes, and is common in research on terrorism, each of these existing datasets are based on different criteria for inclusion and “there is not common definition or coding protocol and thus almost no compatibility” (2000, 13). This creates problems for assessing trends, testing hypotheses, and doing replicable research.

This study takes all of these methodological difficulties into account when formulating criteria for inclusion, coding the data, and understands that the work and conclusions presented here are merely a foundation for further study.

CHALLENGES WITH DATA COLLECTION As with any study of terrorism, there are problems related to data collection especially those of accuracy and accessibility (Silke 2004; Crenshaw 2000). Nearly all data on terrorist attacks used by scholars are from mainstream media and government sources, neither of which is unbiased (Schmid & Jongman 1988, 137-47). Furthermore, scholars of terrorism have tended to engage in very little original data collection and research and demonstrate a “strong reliance on each others’ published work” (137). This study, like the field in general, is limited by the accessibility of credible open source data.

However, with this study, much care has been taken to overcome these limitations as will be outlined below.

While the data set constructed for this study will inevitably be incomplete, in need of regular updates and revisions, and perhaps even questionable in some areas, it will be made publicly available and will provide the foundation for much future quantitative work on state sponsored terrorism. This dataset has several characteristics that will 102 provide an accurate test of the conditions under which state leaders make the decision to sponsor terrorism. First, this dataset began as one where the unit of analysis was the country-year (1968-2001),71 which incorporated a number of variables within and across states, regions, regime types, time periods and leaders. Two sub-datasets were then created and used for the remainder of this study. The first one uses the leader-year (1968-

2001) as the unit of analysis and a second that condensed (summed or averaged depending on appropriateness to the variable) all of the observations to the level of the leader. The multiplicity of perspectives will allow for a comprehensive examination of state sponsorship as a phenomenon.

CASE SELECTION AND THE UNIT OF ANALYSIS In order to properly address the unique nature of state support for terrorism, the dataset initially used the country-year (1968-2001) as the unit of analysis. While state sponsored terrorism certainly existed prior to 1968, there are several methodological reasons for choosing this time frame. First, it was the upsurge in terrorism, and the dramatic increase in sponsorship, following the 1967 Arab-Israeli war72 that brought it to the attention of scholars and policy makers as a phenomenon in and of itself. Second, it captures the bulk of the sponsorship behavior during the Cold War and as much as possible in the post-Cold War era.73

71 And a copy is retained in this format. 72 The rise of Yuri Andropov to the head of Soviet intelligence and the KGB is cited by some as the reason for the dramatic increase in state sponsorship following 1967, not the Six-day War (see Cline and Alexander 1989). 73 Because the Correlates of War Militarized International Disputes data only runs through 2001, I was unable to collect information on the threat environment for leaders after 2001. This will be updated when 103 The countries were selected by cross-referencing the countries listed in the

Correlates of War State System Membership List74 with population numbers from the

World Bank (Population, Total)75 and IISS Military Balance Reports76 to ensure that all states in existence between 1968 and 2001 with a population of 1 million or more were included (N=5132). The unit of analysis for the dataset used in the inferential analysis is the leader-year (i.e. Muammar Qaddafi in 1973, given an observations specific code) means that the data is both panel data (the observations are is nested in the leader) and that it is longitudinal.

Variable L_caseID leader l_code Country Year c_code Name Explanation Unique Leader name Unique Country Observation COW Observation Leader name year country code number number Examples 3382 Muammar 620002 Libya 1973 620 from dataset Qaddafi 3383 Muammar 620002 Libya 1974 620 Qaddafi

Table 4.1: Excerpt from the Leader-year dataset to illustrate coding protocol.

This avoids a dependent variable bias, as discussed above, because we include the information for sponsoring states not only when they are sponsoring, but also before and

the data becomes available, as ideally, there will be an equal number of observations before 1989-1990 as afterward. 74 To be included, “the entity must be a member of the League of Nations or the United Nations, or have a population greater than 500,000 and receive diplomatic missions from two major powers.” Correlates of War Project. 2008. “State System Membership List, v2008.1.” Online, http://correlatesofwar.org. 75 Total population “is based on the de facto definition of population, which counts all residents regardless of legal status or citizenship--except for refugees not permanently settled in the country of asylum, who are generally considered part of the population of their country of origin. The values shown are midyear estimates.” http://data.worldbank.org/indicator/SP.POP.TOTL accessed September 3, 2011. 76 International Institute for Strategic Studies, London volumes 1968-2001. 104 after, as well as including states that do not, and have never, sponsored. This gives the essential “universe” of cases relevant to contemporary state sponsorship.

THE DEPENDENT VARIABLE: STATE SPONSORSHIP OF TERRORISM The biggest challenge with this project was the collection of data on the dependent variable. It may never be possible to conclude definitively that every instance of support has been identified; yet this section will demonstrate that the data is reasonably complete. With the initial assistance of an undergraduate research team who were trained in data collection in the context of an undergraduate course on research in terrorism, insurgencies, and guerilla warfare, the information on the dependent variable was drawn initially from the START terrorist organization profiles, Schmid and Jongman’s (1988) directory of terrorist organizations, and the US State Department’s annual Country

Reports on Terrorism and Patterns of Global Terrorism. The searches gradually expanded to include articles on Lexis-Nexis, peer-reviewed journal articles, books written on specific countries, conflicts, and terrorist groups as well existing datasets such as the

UCDP/PRIO Armed Conflict dataset and Salahyan et al.’s Non State Actor Dataset

(Version 2.7, 2009).

Information was collected on the group name, the type of support--protection, territory, arms, money, training, intelligence, or diplomatic resources--details on that support, and then listed a full bibliographic reference for each source. Despite the questionable nature of data on terrorism in general, a key strength of this dataset is that any case of sponsorship had to be verified by two sources in order to be included in the analysis and all of the source material has been retained and cataloged in the excel 105 version of the dataset allowing for others to reference the sources, verify, and complete additional cases where only one source has so far been located or challenge included cases as new information becomes available.77

Upon completion of the semester and eventual dissolution of the research team, the data collection continued by the author until every reasonably available source had been exhausted and the previously collected data was crosschecked, verified, and corrected if necessary. This research was extensive and a full bibliography is available in

Appendix B. Importantly, the dataset will be made public, with complete bibliographic information, to be peer reviewed and subject to updating.

Qualifying Groups There are often overlaps between terrorist and insurgent groups, which poses conceptual and empirical problems for scholars (see Byman 2005). Indeed, groups do not always use terrorist tactics; rather, they tend to evolve in and out of favoring the use of certain tactics over others making it difficult to categorize a particular group as

“terrorist.” More recently, many scholars, such as San Akca and Bapat, use the generic terms “nonstate actor” or “nonstate armed group” to deal with this overlap. However, supporting groups that specifically perpetrate terrorist attacks entails unique risks for the leader of the sponsoring state. The reprisals, sanctions, censure, and other consequences of sponsoring a group considered to be “terrorist” are significantly higher than for

77 Instances of sponsorship where only one source was found were kept in the dataset and coded as “alleged” support, but not included in the analysis. 106 supporting insurgent groups. This study is concerned specifically with the sponsorship of terrorist groups.

There are several methods that could be used to distinguish terrorist groups from insurgencies for the purpose of this study. First, terrorists are not tied to territory as exclusively as insurgent groups are and it is this feature that makes them a more powerful tool for projecting a state’s power beyond it’s borders, makes them more difficult to retaliate against, and provides a tentative point of delineation between terrorist groups and insurgents (Hoffman 2006). However, determining the extent of territoriality is laden with empirical and operational difficulties. Not the least of which are the changes in transportation and communications technologies in the last few decades, which can make even the most hardened territorially based insurgency appear highly mobile and virtual.

The second approach is to deal solely with the selected target. While both terrorist groups and insurgent actors target military (hard) and nonmilitary (soft) targets, insurgent groups are more likely to focus attention on attacking the adversary governments armed forces, while the terrorist group is more likely to select nonmilitary targets. Thus, using the terrorist groups presented in Schmid and Jongman (2008) and the Terrorist

Organization Profiles (TOP) available from START (Miller & Smarick 2012), a comprehensive list of potential terrorist groups was assembled. This list was then crosschecked with the GTD database, where all of the attacks, targets, fatalities, and causalities for the groups active between 1968-2001 were cataloged.

Unlike the UCDP/PRIO Armed Conflict dataset, which both San Akca and Bapat begin with, and which requires a minimum of 25 battle deaths to be included; for a group 107 to be included as “terrorist” for this study, it had to have been included on either the

START TOP or the GTD, had to have committed at least two attacks against non-military targets,78 over the span of more than one year with no more than 50% of total attacks perpetrated against military targets. These criteria allow for the realities that many, if not most terrorist attacks result in little or no causalities and that a state is more likely to support a group that perpetrates more than one attack and that is around long enough

(more than one year) to allow for logistical and other arrangements to be made to begin supplying the group.

This reduced the original START/Schmid & Jongman list of 856 groups down to

504 “terrorist” groups that were operational between the years of 1968 and 2001.

Importantly, groups were only eligible to be included in the dataset if 50% or more of their attacks were against a nonmilitary target. These groups account for a total of 39,261 attacks (32,888 of them on nonmilitary targets). The list of qualifying terrorist groups is included in Appendix A.

QUALIFYING SUPPORT I did a survey of articles, speeches, and books written by academics, policy makers, and international legal scholars, though almost exclusively American, who have written specifically on state sponsorship of terrorism 1975-2010, this revealed a number of commonalities in what is considered state sponsorship.

78 A non-military target includes attacks against businesses, the food and water supply, the government (general), the government (diplomatic), the police, and private citizens and property. 108 Over 80% of the authors in the works surveyed (N=23) considered the provision of arms, money, and/or assistance with intelligence, logistics, and operations to be state sponsorship. Over 60% also include safe haven/sanctuary and training to that list.

Scholar Inclusion of Types of Support (N=23)

Arms Money Savehaven Training Ideological Intelligence Organizational Govt. Participation

0 5 10 15 20 25

Figure 4.1: Frequency of the types of support included by scholars of state sponsorship in 23 surveyed works.

There are obviously questions as to whether this is consistent across the time periods discussed here or whether there is variation that corresponds to other factors, such as the evolution of international norms. These behaviors, however, correspond with the discussion in Chapter 1 regarding the assets of the state and what resources can be potentially allotted to the terrorist group.

109 Types of Support Given the confluence of available state assets and prevalence in the literature on state sponsorship, this study codes a state leader as supporting terrorism when they allocate one or more of the following resources to a terrorist group.

Protection: The state uses its power and resources to provide protection to terrorists from foreign governments without allocating territory.

a) Safe haven b) Protection from extradition or prosecution c) Transit across territory d) Providing cover via front companies e) Allowing groups to open and maintain political offices.

Access to Territory: The state allocates part of its territory for use by terrorists. a) Training camps b) Bases of operation c) Sanctuary

NOTE: This does not include instances where the state is unable to prevent terrorists from seizing the territory on their own.

Money: The state provides the terrorists access to the financial resources of the state.

a) Financing specific operations b) Providing limited or unlimited funds for the organization/group.

NOTE: This does not include instances where the state does not block individuals from raising money or funding groups on their own.

Arms: The state provides terrorists with weaponry.

110 a) Could range from light weaponry to heavy artillery, rockets, and bomb making materials. b) Includes orchestrating third party arms transfers.

NOTE: Does not include lax supervision of arms transfers on its territory.

Intelligence: The state provides terrorists with access to intelligence presumably gathered by state agencies.

a) Includes logistical support that does not utilize the diplomatic institutions of the state. b) Knowledge transfer from the state to the group.

Training: The state provides organizational and operational “know-how” to the group.

a) Basic and specialized military training. b) Basic and specialized intelligence training. c) Assistance in organizing and institutionalizing. d) Skill transfer from the state to the group. e) Military Advisors

Diplomatic resources: The state provides terrorists with access to its diplomatic institutions and processes.

a) Provision of passports, visas, and other travel documents. b) Confining diplomatic status to terrorists, providing diplomatic immunity to terrorists. c) Use of diplomatic pouch to transport weapons or other material and the use of embassies on foreign soil to provide sanctuary or staging grounds for terrorists.

A final category of "general support" was included in the dataset to cover sources that proclaimed that the state and/or leader provided "support," "military support," or

111 "general support to the terrorist groups. Predominately, this support was explained to be supporting military operations.

Figure 4.2: Binary Categories of Support occurrence in the country-year dataset

The Protection and Territory categories, which could be combined as general category of “safe haven” accounts for 33% of the total incidences (N=1504)79 while those types of support associated with military activities—Training, Arms, and Intelligence— were 30.4% of the total. The provision of Money accounted for 12.8% and the “general support” category was 23.9% of the total. If money is included as a “military” support

79 Note: Because more than one category of support could be observed for each leader-year, the total for this figure is larger than for the “state sponsorship” binary variable. 112 type, then active provision of non-safe haven resources to facilitate military activity accounts for 43.3% of all types of support included in the dataset.80

CODING THE DEPENDENT VARIABLE There are several ways that state sponsorship can be coded. Some rely heavily on categorizing and subcategorizing state sponsorship based on the type and level of support given to the terrorist group (Murphy 1989; Pillar 2004; Byman 2005 & 2008; Shay 2005;

Nacos 2010) or by the level of direct government involvement in the attacks (Stohl &

Lopez 1988; Mikolus 1989) while others code it simply as a binary variable (Bapat

2007).

There is a long history of state sponsored terrorism scholars attempting to develop typologies in order to measure the degree of support. Edward Mickolus developed one of the first sub-category typologies of state sponsorship for use in the ITERATE III dataset of terrorist incidences81 that ranked states by the “degrees of government support/attitudes toward terrorists” (1989, 3). Using 32 observable actions82 that states take in relation to terrorist groups, he developed a rough continuum composed of five general categories that increase in order of state participation: 1) intimidated governments; 2) ideologically supportive regimes; 3) generally facilitative supporters; 4) direct support in incidents by governments; and 5) official participation. At the one end, then, this typology captures

80 This number is as potentially as high as 67% if one takes into account that the bulk of all “general support” is noted to be military support. 81 Covering 1980-1987 82 These state behaviors include, but are not limited, to allowing groups to open offices, safe houses, or otherwise fundraise and recruit within their territory, providing them with money, passports, arms, maps, intelligence, organizational assistance, military-style training, and propaganda or “rhetorical support” (Mickolus 1989; see also Byman 2005; Cline & Alexander 1986). 113 state behavior includes little direct action taken by the government, or “passive” state sponsorship (Byman 2005 & 2008), to direct involvement by agencies and officials of the state in perpetrating the attacks at the other. Murphy (1989) also categorizes state sponsorship by the type of support the state provides, while others differentiate type of sponsorship by whether the state intentionally provides resources or merely allows groups to operate (Pillar 2004), or by the level of involvement of the government in the attacks (Shay 2005; Stohl & Lopez 1988).

Byman, however, argues that the range of state support is more nuanced than merely being comprised of active “sponsors” and passive “enablers” (Pillar 2004), but rather conceptualizes them as strong supporters, weak supporters, lukewarm supporters, antagonistic supporters, passive supporters, and unwilling hosts (2005, 15).83 All of which are categorized based on the sponsoring regimes’ level of commitment to the terrorist group (via state policy), the type of power relationship between the state and group (as state capacity), and the amount of resources that the state makes available (the mix of the two). In later examinations of this variation, Byman re-explores it as subcategories of active sponsorship: direct control by the state, coordination between the group and state, and simple contact between the two, and subcategories of passive sponsorship: knowing toleration, unconcern or ignorance, and incapacity on the part of the state to prevent domestic actors from providing support (2008, 3-4).

However, there are several reasons why categorizing state sponsorship by degree of support is difficult. First, states do not remain consistent in the type or the amount of

83 Bapat 2007 also adopts this typology. 114 support provided to any particular group over time and can provide varying types and degrees of support for different groups during the same period of time. For example,

Syria actively trained, funded, and fielded some terrorist groups as proxies in the past, while actively suppressing the activities of others. Yet, under Bashar al-Assad, Syria now tends to provide only transportation, logistic support, and safe haven for the leadership of groups that it still supports. Likewise, the East German regime provided diplomatic and material support for European terrorist groups during the Cold War, but ceased its activities altogether after the fall of the Berlin Wall, as did Czechoslovakia, Bulgaria and

Hungary. Meanwhile, Qaddafi’s Libya went from being one of the most dangerous direct supporters of terrorism to a state that officially renounced all support for terrorism and began actively jailing and repressing Islamist terrorist groups within Libya.

Second, the risks associated with types of support vary over time and may depend on the development of international norms (Finnemore 1996) or the socialization of new states into the international system (Waltz 1979). Changing international norms may even effect the categorization of behavior as state sponsorship. For instance, in 1986, Libya suffered reprisal attacks from the US for providing diplomatic resources, bomb making materials, money, intelligence, and training to terrorist groups operating in Europe while several other states that “merely harbored” terrorists received little if any attention and certainly no military strikes. Yet, in 1998 the US bombed parts of Afghanistan and

Sudan, not because their regimes were providing al Qaeda with weapons, training, etc. but rather because they were allowing them sanctuary on their territory.

115 These dynamics indicate that the relationships between state sponsors and the sponsored terrorist groups are not static. The level and type of support may vary by the intention and policy goals of the state, the domestic constraints placed on the resources available to the regime, the normative pressures of the international community or it could simply be determined by the resources the group already has access to (see for example Byman and Kreps 2010).84 This study takes the position that the level and type of support provided to terrorist groups by state leaders is variable, and is worthy of further inquiry and empirical investigation, but is not appropriate as an ordinal measure of regime behavior until more is known about its variance.

Figure 4.3: Binary State Sponsorship Variable breakdown in the data

84 See Shoemaker & Spanier (1984) for a parallel examination of patron-client state relationships. 116 Taking these issues into consideration, the dependent variable for this study: state sponsorship of terrorist groups, is coded as a binary variable for each observation (leader- year) with the values of (1=indication of state support; 0=otherwise). This resulted in approximately 16% of the observations from the leader-level dataset being coded as

“sponsorship.”85

THE EXCLUSION OF “PASSIVE” SUPPORT AND PRIVATE FINANCING A quick detour into the examination of “passive” support is appropriate here. An additional complication in classifying state sponsorship arises from the fact that scholars and policy makers now consider the unintentional provision of sanctuary and the inability or unwillingness to suppress terrorist groups a type of, “passive,” state sponsorship

(Byman 2005, 2008; Pillar 2004). Indeed, there is a widespread concern that the weaker states of the post-Cold War era are increasingly providing sanctuary, and thus operational capacity, to terrorist groups (Byman 2005, 2008; Piazza 2008; Cronin 2002).

The term “state sponsorship” has stretched to include passive support by weak states, especially in the post-Cold War era. It has also stretched to include situations where states do not prevent private individuals and institutions from financing groups tied to terrorism. For example, the Saudi government continues to tolerate Saudi individuals and charities that finance the radical Salafi madrassas in Pakistan, which are increasingly linked to attacks in Afghanistan (Moghadam 2009; see also Levitt 2002). The Pakistani government struggles to balance the domestic pressures for tolerating Islamic activism

85 The individual categories of state sponsorship were also coded in this same manner, “1” if there was an indication of sponsorship, “0” otherwise. Importantly, only the cases where two independent sources indicated sponsorship were included in the analysis. 117 with the growing threat to its regime and with the external pressure from the US to continue pursuing suspected members of al-Qaeda in the northern tribal areas. Some scholars have argued that these cases of tolerance and sanctuary provision exemplify larger, global patterns of change in state-terrorist relations (Byman 2008).

Indeed, some states that are considered to “sponsor” terrorism today do so only in the sense that they are unwilling or unable to suppress terrorist activities emanating from within their borders (Byman 2008; Piazza 2008) often because the domestic cost of shutting them down could well be the survival of the regime itself. This study argues that passive sponsorship in these cases may not be illustrative of different types of the same phenomenon. It could be the case that the foreign policy of actively providing terrorists access to state resources is not a comparable behavior to that of being unable to eject a group from its territory.86 The research and policy implications of treating them as opposite ends of a continuum of state support have not been adequately examined.

Therefore, this study focuses on the active provision of state controlled resources, including territory, to terrorist groups and leaves an empirical investigation into the causes of passive sponsorship and its relationship with active support for future research.

Given the fact that the relationships between states and terrorist groups are highly dynamic (Byman 2005) and that the international litmus for what constitutes state sponsorship also changes, it is difficult, if not impossible to accurately assign a static

86 See San Akca 2009 for an attempt to develop separate selection processes for groups deciding where to locate their bases—more likely to occur in weak or failed states—and states deciding whether to provide intentional support. 118 “degree” of state support and thus the binary coding of the dependent variable is the most appropriate until the dynamics of state-terrorist group relationships are better understood.

INDEPENDENT VARIABLES The independent variables for this study are taken directly from the state sponsorship, civil-military relations, and foreign policy literatures explored above and are designed to test each of the hypotheses presented in Chapter 3. This section outlines the independent variables of regime type, threat environment, and military expenditure and discusses how they have been measured and operationalized.

Regime Type The method of analysis, as will be explained below, is a time-series logistic regression. Given that the dependent variable is binary, it is important for all of the other variables to be as close to that scale as possible. With that in mind, and given its widespread use, the regime type variable was measured using data drawn from the Polity

IV Project data. However, rather than relying on just the Polity variable, this study utilized the Polity2 variable, which is an adaptation of the Polity score that attempts to fill in some gaps where data is missing, thus allowing for more cases to be included that would omitted if using the regular Polity variable (Polity IV Project: Dataset Users’

Manual). The Polity2 scores, like the regular polity variable, run from -10 to +10 with the more authoritarian regimes in the negative and the more democratic regimes in the positive.87

87 Missing values were left blank and thus, were omitted from the regression. 119 Initially, the regime variable was broken into the four categories used by the

Polity IV project, that is, Authoritarian regimes (with values of -10 to -6); Closed

Anocracies (-5 to 0); Open Anocracies (0 to 5); and Democracies (6 to 10).88 However, for logistic analysis, simplicity is key and categorical/ordinal independent variables are difficult to interpret. Since the existing literature on state sponsors argues that autocratic leaders are more likely to sponsor terrorism, the regime type variable was converted into a binary “autocracy” variable.

Furthermore, in order to stabilize89 the Polity2 score, I followed Bueno de

Mesquita et al.’s (2003, 137) protocol where each score (-10 through 10) was added to 10 and then divided by 20 in order to give it a value between 0 and 1. Those scores that were closer to 0 were more autocratic while those closer to 1 were more democratic. Thus, for the regression analysis, the regime type was binary coded as autocracy (1) when it had the Polity2 adjusted scores of 0 to 0.5 (0= Otherwise). A “democracy”90 variable was created for some comparison in the descriptive data (1= Polity2 adjusted scores 0.51 to 1;

0= Otherwise), specifically in the leader-condensed dataset.91

88 Note, Polity IV actually differentiates between democracy (6 o 9) and full democracy (with a score of 10), however, for simplicity this study combined the two into “democracy 89 Specifically, to get rid of the negative values and make it easier for STATA to handle the data. 90 While some of the regimes that were coded as “democratic” were, in fact, open Anocracies, this does not affect the inferential analysis because only the more autocratic (closed Anocracy and authoritarian) regimes were regressed against the dependent variable. 91 Where the regime type was coded based on the average Polity2 score over the course of the leader’s tenure. 120 THREAT ENVIRONMENT The second set of independent variables relate to the threat environment faced by the leader. The literature and anecdotal accounts strongly suggest that external conflict contributes to state sponsorship, and the empirical tests confirm this. The Theory of

Leadership Survival presented here, however, anticipates that internal threat of coup d’état by the military also contributes to a leader’s decision to sponsor terrorist groups. It puts forth the argument that the combined threat environment the leader faces is a key determinant of state sponsorship.

Therefore, a series of binary variables were created to capture the different discrete combinations of internal & external threat that leaders face: High External &

High Internal (HEHI), High External & Low Internal (HELI), Low External & High

Internal (LEHI), and Low External & Low Internal (LELI). Thus, those cases with a high threat environment only (HELI) will be examined as separate from those that have both a high external and internal threat environment (HEHI).

To measure external threats to leaders and their regimes, this study relied upon the

Militarized International Disputes dataset (v3.10) available from the Correlates of War

Project.92 This data is valuable for estimating the occurrence of state sponsorship because it relies on not only instances of full-scale war, but also records when states threaten, display, or use force against another state (Ghosn et al. 2004). One caution regarding the

MID variable is that, since the data was originally coded for country-year and then converted to leader-year, there are some MIDs that may be accounted for twice. The

92 This data was only available through 2001, which provided the upper limit of the dataset. 121 conversion of the MID variable to a binary (1= one or more MIDs; 0=otherwise) rather than a count variable, however, should minimize any risk of over-representation of MIDs.

The internal threat variable, on the other hand, was directly coded to the leader- year. The number of coup plots, coup attempts, and successful coups, along with the exact dates of these coup occurrences are listed in the Center for Systemic Peace Coups d’état data (Marshall & Marshall 2010), allowing each event to be precisely placed with the correct leader. This precision was key given that many leaders came to power by a coup d’état or were threatened by coup but retained their position. The internal threat binary variable (1= threat of coup; 0=otherwise) was then constructed using two measures: first, the leader was coded as facing threat of coup if the previous regime faced a coup threat, including if the current leader came to power via coup, and second, if the current leader himself faced a coup attempt.

In order to examine the effect of the possible combinations of the external (MIDs) and internal (coup d’état) threat environment, then, the four binary variables were constructed (1,1; 1,0; 0,1; 0,0) for both the leader-year and leader condensed datasets in order to see if the descriptive data lent any credibility to the theory and to determine if it was worth continuing on to the inferential analysis.

One area of challenge outright was the question of whether the variation in internal threat level makes a difference if the high external threat is consistently significant, in other words, why not just look at HE by itself? This study does not discount the fact that the external threat is a major explanatory variable. However, as the descriptive data below illustrate, both leader-year cases and leaders facing combined high 122 internal and external threat environments (HEHI) have a much higher rate of sponsorship, represent a larger number of key cases, and represent a larger proportion of sponsorship cases to total cases, than do cases and leaders facing the high external threat alone (HELI)

Total Threat Total State Threat State Sponsorship Leader-years Environment Sponsorship Environment Cases as % of Threat (N=5132) Cases Cases Cases as % of Environment Total HEHI 690 285 13.4 41.3 HELI 987 315 19.2 31.9 LEHI 1351 138 26.3 10.2 LELI 2110 192 41.1 9.1 Leaders (N=967) HEHI 177 66 18.3 37.2 HELI 353 105 36.5 29.7 LEHI 148 9 15.3 6.1 LELI 289 10 29.9 3.4

Table 4.2: Threat Environments that leaders face and the occurrence of state sponsorship.

The data, then, indicate that both the High External & High Internal (HEHI) and the High External & Low Internal (HELI) threat environments had levels of sponsorship well above those in the other two categories. Indeed, the HEHI threat environment was higher than HELI in both the leader-year and when condensed to the level of the leader.

While only 18.3% of state leaders faced an HEHI threat environment, 37.2% of those leaders sponsored at least one terrorist group. There are indications, then, that the threat

123 environment faced by the leader has a noticeable relationship with the decision to sponsor terrorism and that HEHI threat environments are the most acute.

Military Expenditure as % of GDP The final independent variable that this study includes is the amount of money that is earmarked for the military. Acknowledging that resources are scarce, and that political processes within the state affect the distribution, the theory predicts that sponsoring leaders will be under pressure to allocate larger shares of these resources to the military. This pressure is a function of both the high external threat environment that requires preparedness and mobilization as well as the internal pressure to provide private goods to maintain the support of the military for the regime. While existing literature on state sponsorship does not mention military expenditure explicitly, it does imply that, as a component of military strength, lower expenditures would be expected given that sponsorship is thought to be a way to supplement weak conventional armies.

Because state GDPs and budgets are so incredibly disparate, this study attempts to equalize them by using yearly military expenditure as a % of the yearly GDP

(gdp_mil1995).93 The military expenditure variable was obtained from the Correlates of

War National Military Capabilities Index, as were the military expenditure figures. In order to account for as many years as possible, when there was missing data, the differences of the next several years were averaged and subtracted from the subsequent year, or added to the prior. While this does not provide an exact accounting of surges and

93 All figures were converted to 1995 $US in order to maintain standardized comparability across regimes and time. 124 lulls in military expenditure, the large N (5132) and the nearly complete availability of this data (cross-checked with data from the World Bank and IISS to ensure that the estimates were not biasing the predictor) provides a relatively high level of confidence in the estimated values. The gdp_mil1995 variable was then computed by dividing the military expenditure (in standardized 1995 $US) by the GDP (also in standardized 1995

$US).94

The regime type, threat environment, and military expenditure variables will be used to explicitly test the three hypotheses presented in Chapter 3, yet there are two other variables that need to be controlled for in the analysis: the internal threat to the regime from internal war and the security structure of the Cold War.

CONTROL VARIABLES Two control variables, internal war (int_war) and Cold War (cold_war), were also introduced into the model. First, the internal war variable95 (int_war) accounts for the threat posed by both ethnic and revolutionary war96 to the regime. Much of the state sponsorship literature understands legitimacy of the leader to be a factor in a leader’s decision to sponsor terrorism. Internal militarized threats, in the form of ethnic and revolutionary war, against the ruling regime are reasonably partial indicators of this legitimacy. A binary variable for internal war (int_war) was introduced to capture the

94 Using a percentage rather than the raw dollar values also has the advantage of returning more accurate coefficients in the logistic analysis. 95 Sourced from Political Instability Task Force (PITF) Ethnic Wars 2009 & PITF Revolutionary Wars 2009 data available from http://globalpolicy.gmu.edu/political-instability-task-force-home/pitf-problem- set-annual-data/ accessed December 10, 2012. 96 If the state experienced revolutionary war and/or ethnic war during the observation year it was coded 1=yes; 0= otherwise. 125 threat emerging from both revolutionary insurgent war (those intending to overthrow the incumbent government) as well as ethnic/separatist war.97 Furthermore, a series of interactive variables were explored to determine if there was a multiplying effect of internal war on the internal threat faced by regime from the military.98

Second, a control variable for the Cold War was introduced to help test the commonly held assumption that state sponsorship was most prominent during the Cold

War (Byman 2008). The Cold War variable is expected to be significant, though not necessarily for the structural reasons supposed by other scholars. While this research design is not equipped to answer the question of why there is variability between the

Cold War and post-Cold War eras, the theory would suggest that the decrease in HELI and HEHI threat environments would account for a large portion of this decline.

Coding for the years 1989-1991 are “fuzzy” in respect to belonging to the Cold

War or post-Cold War era, yet following the fall of the Berlin wall the Soviets (and other bloc sponsors) severed many foreign ties and turned radically inward to deal with domestic issues signaling a shift in the structure in 1989. Therefore, the Cold War variable (cold_war) was coded “1” for the years 1968-1989, while 1990-2001 were coded

“0.” Due to concern that the inclusion of the Superpowers during the Cold War in the regression was skewing the military expenditure variable or adding more weight to sponsorship under the circumstances of the HELI threat environment, a filter was used to

97 An interactive variable of the threat environment and internal war was also examined in each of the models to see if internal war had an amplifying effect on the threat environment, especially in HI situations. This interactive variable was insignificant in all models. 98 Interactive variables of int_war*HELI, int_war*HEHI, int_war*LEHI, and int_war*LELI were also regressed in an iteration of Models 1-4 below. None of the interactive variables were statistically significant. 126 exclude the Superpowers in an addendum to the fully fitted model (and the results presented in Model 7).

The dependent, independent, and control variables have been explored at length here and the method of analysis and results will be explored in the next section.

METHODS OF ANALYSIS AND REGRESSION TABLES The dependent variable of state sponsorship in this study is binary, requiring a logistic analysis, and given the longitudinal and panel characteristics of the data, the appropriate method for examining the relationship between dependent and independent variables is a cross-section time-series logistic regression99 (Long & Freese 2006) using an adaptive quadrature (Hosmer & Lemeshow 2000). These models use random rather than fixed effects, which provide more accurate indicators of the parameters with a large-

N, but a small (or varied) T (Maddala 1987), and allow not only for variation within the individual s (leaders), but also for the inclusion of some variables that are different across leaders but that remain fairly invariant for a single leader over time, such as military expenditure as percent of GDP (Yaffee 2003).

Multiple models were fit in order to capture the threat environments in isolation and in combination. The first series of models (Models 1-4) fit the models using the threat environment variables as series of four binary variables that were each examined independently. The second series of models (Models 5-7) were run using Model 2 (HEHI only) as a baseline, since it was the variable of interest for this study and it was important to determine if adding additional threat environments altered its significance.

99 xtlogit in STATA 127 For Models 5 & 6, one threat environment variable was added each iteration to determine if the inclusion of an additional threat environment made a difference in the significance and to provide controls on the threat environments themselves. Model 6 is the fully fit model, while Model 7 is fully fit, but excludes the Superpowers to ensure their mammoth military GDPs weren’t biasing the model.

Model 1 (HELI only)100 Model 1 looked at the High External & Low Internal (HELI) model in isolation.

This is the model that most closely aligns with the perspectives of most state sponsorship scholars today. As Figure 4.3 below shows, the data indicates that the HELI threat environment is indeed significant at the p>0.001 level of significance.101

Autocratic leaders were also significantly likely to support terrorist groups in an

HELI threat environment, as were leaders who faced a threat of internal war. Military expenditure was also highly significant at the p>0.001 level of significance. As expected, the Cold War dummy was significant, although unexpectedly, it had a smaller coefficient than the other IVs and the other control variable of internal war. Internal war, in fact, was consistently significant across all models and is likely to serve as an additional internal threat to the regime. In sum, the model indicates that there is empirical support for the

100 Recall that the binary variable for the threat environment of High External & Low Internal (HELI) threat was constructed by coding a 1 if the variable for High External threat (MID) was 1 and the threat of coup was 0. Table 4.2 shows that this threat environment clearly affects the largest number of leaders in the dataset (353 leaders or 36.5%) and a very high number of observations 987, or just over 19%). 101 The STATA command xtlogit b_sst HELI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67) 128 argument that leaders facing an HELI (external only) threat environment are quite likely to provide support to terrorist groups.

129 N=5132 DV: State Sponsorship Model 1 Model 2 Model 3 Model 4 Threat HELI Environment ß 1.241*** sd (.2403) OR 3.46 HEHI 1.162*** (.2719) 3.20 LEHI -.8890*** (.2511) 0.41 LELI -1.276*** (.2373) 0.28 Regime Type Autocracy 1.506*** 1.216*** 1.432*** 1.257*** (.3639) (.3607) (.3651) (.3579) 4.51 3.37 4.19 3.52 Military mil_gdp1995 2.871* 2.935* 2.879* 2.945* Expenditure as (1.261) (1.284) (1.278) (1.271) % of GDP 17.66 18.83 17.79 19.00 Internal War int_war 1.170*** 1.019*** 1.060*** 1.159*** (.2863) (.2906) (.2907) (.2846) 3.22 2.77 2.89 3.19 Cold War cold_war .8432*** .7141** .6968** .8798*** (.2450) (.2450) (.2464) (.2443) 2.32 2.04 2.01 2.41 Constant -8.054*** -7.756*** -7.506*** -7.208*** (.5045) (.4929) (.4949) (.5049) .0003 .0004 .0005 .0007 Observations Obs 4959 4959 4959 4959 Groups Groups 949 949 949 949 Statistics log likelihood -1116.46 -1120.67 -1123.62 -1115.13 ln_sig2u 3.336 3.385 3.381 3.352 sigma_u 5.274 5.433 5.423 5.345 Rho .8942 .8997 .8994 .8968 wald ch2(5) 72.70 65.12 58.13 76.17

Table 4.3: Regression Models 1-4 fit with each threat environment examined separately. 130 MODEL 2 (HEHI ONLY)102 Very similar to Model 1 that examined leaders facing the HELI threat environment, Model 2 shows the predicted direction and significance of the HEHI variable. When examined in isolation, the odds-ratios for the HEHI103 variable indicate that HEHI leaders have a 3.2% higher odds of providing support for a terrorist group than do non-HEHI leaders (and the variable is significant to the p>0.001 level of significance).

Autocracy remains significant (p>0.001), though takes on a smaller coefficient than in

Model 1, as do the coefficients for the variables of internal war and Cold War. At the same time the coefficient for the variable examining military expenditure as a percentage of GPD increases. This lends tentative quantitative support for the hypothesis that as the military becomes better funded and equipped in an HEHI environment, it poses a greater threat to the leader via coup d’état prompting them to sponsor.

While there are still multiple examinations to make, there are indications that there is a level of support for all posited hypotheses: 1) that a leader facing an HEHI threat environment will be most likely to sponsor terrorism; 2) that authoritarian leaders, who rely on the military more to remain in power are more likely to sponsor terrorism; and 3) that as military spending increases, leaders are more likely to sponsor terrorist groups.

102 The binary variable of High External & High Internal (HEHI) was created for cases that had a value of 1 for the external variable (MIDs) and a value of 1 for the internal threat variable (coup threat). Table 4.1 shows that of all the threat environments, the highest percentage of cases (41.3%) and leaders (37.2%) facing this threat environment were sponsors. The High External & High Internal cell accounts for the threat environment of 13.4% of all leader-year observations and the threat environment for 18.3% of all leaders. 103 The STATA command xtlogit b_sst HEHI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67)

131 Model 3 (LEHI Only)104 The Theory of Leadership Survival expects that leaders facing a low external threat (MID coded 0) and high internal threat (coup threat coded 1) will lack the incentives or pressures to sponsor terrorism because support for external groups will not assist the leader in alleviating the internal threat. Thus, it is expected that state sponsorship will not occur in the context of this threat environment.

Model 3, indeed, shows that internal threat by itself does not account for state sponsorship. The LEHI threat environment is negative and significant (p>0.001), indicating that, as predicted, leaders in the LEHI threat environment are unlikely to sponsor terrorism. However, the coefficients and significance of the other variables have changed very little. The coefficient on the military expenditure variable dropped slightly from the HEHI model, while the regime type and internal war coefficients increased. The

Cold War variable remained significant, though the coefficient dropped there as well.

Given these results, it is appropriate to take another look at regime type. Although the regime distribution is very similar to that seen in the HEHI category105 (i.e. autocratic regimes are over-represented in both), the rate of sponsorship for LEHI leaders is very low (only 6.1%) while for HEHI leaders it was 37.2%. Thus, despite the fact that there are nearly the same number of autocratic leaders in LEHI (111) and HEHI (141), only

4.1% of the LEHI autocratic leaders were sponsors compared with 30.5% of HEHI autocratic leaders.

104 STATA command xtlogit b_sst LEHI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67) 105 Approximately 75% of all LEHI leaders were autocratic while for HEHI leaders the number was 80.2% 132

Table 4.4: Leader Sponsorship and Threat Environment by Regime Type

In other words, regime type alone is not able to account for the leader decision to sponsor terrorism. The attribution of state sponsorship to autocracy is perhaps, in part, better understood as a function of threat environment.

Model 4 (LELI Only)106 The cases coded LELI have values of 0 (zero) for both the external threat environment (MIDs) and the internal threat environment (coup d’état) variables. Once again, under these circumstances, the Theory of Leadership Survival expects that leaders will not gain any utility from, and therefore will not, sponsor terrorist groups. Table 4.3 above shows that, in line with the predictions of the theory, virtually no leaders facing a low internal & low external threat environment sponsored terrorism (10 or 3.4% of all

106 STATA command xtlogit b_sst LELI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67) 133 LELI facing leaders) despite the fact that 289 (29.9%) of all leaders faced this threat environment.

Model 4, which examines LELI cases in isolation, indicates that as predicted, leaders in a LELI threat environment were unlikely to sponsor terrorism (negative and at the p<0.001 level of significance). While regime type, military expenditure and internal war were, once again, all significant.

These initial, isolated, models show tentative support for the Theory of

Leadership Survival. The next step is to examine the variables in combination to examine the effects of the threat environments while the others are held constant.

Model 5 (HELI & HEHI Only)107 In the remaining models (Models 5-7), one of the threat environment categories needed to be omitted in order to prevent perfect collinearity among the threat environment variables (Long & Freese 2006). Since the descriptive data indicates that only 1% of leaders both face the Low External & Low Internal threat environment and sponsor terrorism, and the regression table for Model 4 demonstrates a strong significant and negative relationship between this threat environment and state sponsorship, there was no theoretical reason to retain it. The LELI threat variable was the most logical to omit from this portion of the analysis.

The purpose of Model 5 was to determine if the HEHI threat variable remained significant when the model was fit with the HELI threat held constant. This study

107 STATA command xtlogit b_sst HELI HEHI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67) 134 acknowledges that sponsorship occurs in both the HELI and HEHI threat conditions, but the Theory of Leadership Survival is most interested in the specifics of the HEHI threat environment. In this, still incomplete, model, both threat environments are highly significant (to the p>0.001 level of significance), and the coefficient on the HELI variable is slightly higher, yet the coefficients on both variables are higher than in the isolated models. The coefficient on the internal war variable is higher than in the HEHI only model, but lower than in the HELI model and overall, the variable coefficients tended to average between Models 1 and 2.

135 N=5132 DV: State Sponsorship Model 5 Model 6 Model 7 Threat Environment HELI 1.401*** 1.357*** ß 1.325*** (.2515) (.2525) sd (.2407) 4.06 3.89 OR 3.76 HEHI 1.266*** 1.648*** 1.674*** (.2710) (.4376) (.4382) 3.55 5.20 5.34 LEHI .4538 .4816 (.4073) (.4079) 1.57 1.62 Regime Type Autocracy 1.322*** 1.250*** 1.317*** (.3588) (.3624) (.3648) 3.75 3.49 3.73 Military Expenditure as % of mil_gdp1995 2.973* 2.994* 2.999* GDP (1.264) (1.267) (1.267) 19.56 19.97 20.06 Internal War int_war 1.054*** 1.059*** 1.082*** (.2899) (.2892) (.2897) 2.87 2.88 2.95 Cold War cold_war .7966*** .8160*** .8006*** (.2455) (.2456) (.2465) 2.22 2.26 2.23 Constant -7.966*** -8.138*** -8.229*** (.5099) (.5387) (.2465) .0003 .0003 .0003 Observations Obs 4959 4959 4924

Table 4.5: Fully fit threat environment Models 5, 6 and 7 (that accounts for the Superpowers).

The most interesting feature of this model is that the military expenditure coefficient is higher than in any model presented so far. This result indicates that in either

136 HEHI or HELI threat environments, the percentage of GDP spent on the military has a very strong relationship to a leader’s sponsorship of terrorist groups. Once again calling into question not the theory, necessarily, but the sensitivity of the metric for the threat environment.108

Model 6 (HELI, HEHI, & LEHI) & Model 7 (HELI, HEHI, & LEHI controlled for superpowers).109 Finally, these fully fitted models added the LEHI threat environment to test whether the HEHI variable remained significant. In both models, HELI and HEHI did remain significant, while LEHI dropped from significance.110 In Model 6, the coefficient for HEHI facing leaders actually grew in relation to HELI leaders, indicating that when all of the threat environments are held constant, leaders in a HEHI threat environment are more likely than the others to sponsor terrorist groups. This effect was amplified when the superpowers were filtered out of the analysis. In all, the values on autocracy, military expenditure, and internal war remained positive and significant, as did the Cold War control.

Given that there were nearly twice as many cases in the dataset for the Cold War years 1968-1989 as in the post-Cold War years 1990-2001, the effects of this control variable cannot be reliably interpreted, but the information can be kept in mind for future

108 Specifically the policy divergence between the civilian and military leadership. 109 SATA commands Model 5: xtlogit b_sst HELI HEHI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67); Model 6: xtlogit b_sst HELI HEHI LEHI autocracy mil_exp1995 int_war cold_war, intmethod (aghermite) intpoint (67); and Model 7: xtlogit b_sst HELI HEHI LEHI autocracy mil_exp1995 int_war cold_war if superpowers=0, intmethod (aghermite) intpoint (67). 110 These same models were fit using the LELI variable to see if the omission of it rather than the LEHI variable made a difference in the results. While the value of the coefficients changed, the direction and significance remained the same and the HEHI variable retained a higher coefficient than the HELI one. 137 research when the dataset can be updated so that there are a more equal number of years on either side of 1989-1990. Another explanation regarding the influence of the Cold War is that both autocratic regimes and HEHI facing leaders were more prevalent during the

Cold War. Once more data becomes available on the post-Cold War era, these assertions should all be retested.

Most importantly, for the years 1968-2001, the time series logistic analysis of the data indicates that the three hypotheses posited by the Theory of Leadership Survival have been, at least, tentatively supported by empirical evidence.

RELEVANCE AND PREDICTIVE VALUE The empirical results from this analysis indicate that the data tentatively supports the hypotheses that autocratic leaders are most likely to provide support for terrorist groups when they face a HEHI threat environment and when military expenditure is increased. In the fully fitted model, the prediction that autocratic leaders would be more likely to sponsor received support while the relationship between military expenditure (as a percentage of GDP) and state sponsorship was also consistent across models. While the finding that autocratic regimes support terrorist groups is not surprising, the relationship between military expenditure and sponsorship is one that has not been uncovered by other researchers nor is it one that can be accounted for by other theories. The Theory of

Leadership Survival, however, provides a framework for understanding the relationship between military expenditure, regime type and threat environment faced by the leader on the propensity of that leader to sponsor terrorist groups.

138 Overall, the predictions of the Theory of Leadership Survival were supported.

One of the most important aspects of a theory, other than internal cohesion and empirical validity, however, is its predictive and explanatory capabilities. When known state sponsors are examined through the lens of the Leadership Survival Theory, we can account for the overwhelming majority of the most prominent sponsorship cases.

Leader Country HE HI Sponsor Predict Qaddafi Libya YES YES YES YES Saddam Hussein Iraq YES YES YES YES Khomeini Iran YES YES YES YES Hafez al-Assad Syria YES YES YES YES Omar al-Bashir Sudan YES YES YES YES Castro Cuba YES NO YES NO Chavez Venezuela YES YES YES YES

Table 4.6: Threat Environment for Leaders who’s Countries have been on US State Sponsored Terrorism List and Theory Prediction

We can also use the information gleaned from this study to organize known leaders within threat category “bins” to not only help gain deeper insight to sponsorship in general and motivations of all known sponsoring leaders, but also to keep an eye on potential sponsors.

139 Internal HIGH LOW

HEHI: HELI:

Khomeini, Iran Reagan, USA HIGH Hussein, Iraq Brezhnev, USSR Qaddafi, Libya Castro, Cuba Assad, Syria Afewerki, Eritrea

External LEHI: LELI:

Compaore, Burkina Faso Tun Abdul Razak bin Hussein, LOW Wajid, Bangladesh Malaysia

Daoud, Afghanistan Havel, Czechoslovakia Amos Sawyer, Liberia Rahman, Bangladesh

Table 4.7: Threat Environments and a Sample of Sponsoring Leaders

Indeed, the case studies explored in the following chapters are drawn from these threat environment bins in order to dig deeper into the causal mechanisms, add depth to the data, and examine outliers.

CONCLUSION The results of the cross-section time-series analyses and the multitude of descriptive data indicate that there is empirical support for the hypotheses that the HEHI threat environment is a significant predictor of state sponsorship. While logistic analysis and resulting coefficients cannot predict in a linear fashion (as they must be converted to

140 odds ratios and even then are only cautiously interpreted), enough quantitative evidence supporting the Theory of Leadership Survival exists to allow the study to move to qualitative case studies that will examine the causal mechanisms in depth, examine outliers, and provide a new account of the reasons that prominent leaders such as

Ayatollah Khomeini, Saddam Hussein, and Mummar Qaddafi made the decision to sponsor terrorist groups.

141 Chapter 5: Conforming Cases This chapter will examine several well-known and often studied cases of state sponsorship (Khomeini’s Iran, Saddam Hussein’s Iraq, and Qaddafi’s Libya) that also appear to conform to the Theory of Leadership Survival. The narratives will reframe the cases in terms of the theory in order to see if there is support from the historical record and whether the hypotheses advanced in Chapter 3 and tested in Chapter 4 are at work in these prominent cases of state sponsorship. Specifically, this chapter will look more deeply at the causal mechanisms, examine the high levels of international and domestic pressures that the leaders face, and explore whether the use of terrorist groups helped the leaders to respond to their external threats and maintain credibility within their ruling coalitions through achieving policy objectives that were vital to their rule while balancing the acute threat posed by their conventional militaries. The results of the quantitative analysis in the previous chapter demonstrated that there was significant statistical support for the main tenets of this theory: that leaders facing internal and external threats to their rule simultaneously were most likely to sponsor terrorism, that autocratic leaders were the most likely sponsors, and that the higher the percentage of the GDP that the leaders diverted to military expenditure, the more likely they were to sponsor terrorist groups.

The case analysis below suggests that the Theory of Leadership Survival enjoys support from the historical record, at least in the case of some of the most prominent sponsors.

The Theory of Leadership Survival predicts that leaders facing High External-

High Internal (HEHI) threat environments are most likely to provide exceptionally high level of sponsorship, are most likely to sponsor them intensively, and for longer periods 142 of time given that their regime becomes dependent on this sponsorship in order to secure their place in power. This quantitative portion of this study was specifically developed to test state sponsorship in the HEHI threat environment, thus, it is the key threat environment for the theory to explore in the case studies as well. The number of notorious sponsorship cases represented in the extreme HEHI category (see Table 4.7) indicates that there is indeed an association between sponsorship and threat response.

However, rather than simply selecting a handful of HEHI sponsors to examine at random, a second methodological choice for case selection was based on the length of time a leader was in power, and a state sponsor. The group of cases eligible for selection was drawn from a subsection of leaders who were in power for 10 years or more with a significant number (more than half) of those observation years including state sponsorship. These criteria were chosen for two reasons: first, the data seem to strongly indicate that there is a relationship between longevity of rulers and sponsorship--although the directionality of this relationship is not yet clear. Second, these criteria help sort the minor, temporary, or accidental sponsors from those who have adopted the sponsorship of terrorist groups as a policy tool.

These criteria resulted in the selection of some of the best-known cases of sponsorship, Qaddafi’s Libya, Khomeini’s Iran, and Saddam Hussein’s Iraq. These cases will be reexamined through the framework of the Theory of Leadership Survival, which this study argues will provide a better accounting of the external and internal threats these leaders faced and provide some anecdotal evidence in support of the hypotheses regarding how their support for terrorist groups helped them to keep these threats to their 143 rule in check. Indeed, the Theory of Leadership Survival provides exceptional leverage in these cases.

One caveat moving forward is that while a tremendous amount of research has been on the international and strategic effects of state sponsorship, studies on the domestic impact of this support on sponsoring states are virtually nonexistent (for a major exception see Byman 2005).111 Thus, while it is unlikely that this study will uncover a

“smoking gun,” the intention is to draw clear lines between external threat, internal threat of coup, and the security outcomes of state sponsorship for the leader.

The remainder of the chapter takes each of these cases in turn—Khomeini’s Iran,

Saddam Hussein’s Iraq, and Qaddafi’s Libya--and gives a brief overview of the source and dynamics of their external and internal threat environments. It then asks whether the leader utilized terrorist groups in response to these onmi-threats in the ways anticipated by the theory. Specifically, did they use terrorist groups to decrease the external threat while circumventing the military, build up alternative force structures (especially the intelligence services) that worked directly with the terrorist groups but did not fall within the military chain of command, and did they use these groups to achieve policy goals that were vital to demonstrate the competency of the leader to the remainder of his coalition and ensure the continuation of his rule? This case analysis below provides considerable evidence that this is in fact how things worked.

111 However, he only addresses, in any depth, the domestic security concerns for the cases of Syria and Pakistan. Furthermore, these security concerns focus on societal groups and pressure for legitimacy through “aid to (ethnic) kin” by the leader and does not comment on the role of the military as power brokers in domestic politics. 144 AYATOLLAH KHOMEINI’S IRAN Iran is widely considered the most active and dangerous contemporary sponsor of international terrorism. It was designated a state sponsor by the US Department of State in 1984 and one of the most recent Country Reports on Terrorism (2011) accuses Iran of having, “increased its terrorist-related activity,” including the attempted assassination of a

Saudi Ambassador, “likely in an effort to exploit the uncertain political conditions resulting from the Arab Spring, as well as in response to perceived increasing external pressure on Tehran” (US Department of State, Country Reports on Terrorism 2011). Iran, then, in addition providing insight into period-specific instances of sponsorship, is also a case that helps to illustrate its institutionalization.

Leaders of Iran have sponsored terrorist groups for over three decades, despite having transitioned to new leadership, and having been subject to multiple levels of sanctions. Although the Islamic Republic of Iran has, fairly consistently since the early

1980s, provided training, arms, funds, intelligence and diplomatic support to numerous groups that perpetrate terrorist attacks--the most famous of which is Hezbollah--this section will look specifically at the threat environment and sponsorship behavior of

Ayatollah Khomeini, a man largely thought responsible for its inception.

Ayatollah Khomeini came to power following the social revolution in 1979 that ousted the Shah of Iran and instituted a new era of Islamic rule. While Khomeini was not the first Iranian leader to sponsor terrorist groups (San Akca 2009; Salahyan et al 2009), he did structurally change the Iranian system in response to existing internal and external threats, which created a situation where the survival of the leadership became dependent

145 on this policy. Khomeini’s newly consolidating leadership faced extreme internal and external threats in the form of challenges from the military and a conventional war with

Iraq. In this context, Khomeini made many decisions that allowed him to stabilize his regime and minimize the threat posed by the military while maintaining the ability of the state to avoid defeat in the war with Iraq. This study will argue that sponsorship of terrorist groups was one of them.

High External Threat Environment for Iran The story of Iranian sponsorship is normally told in the context of the adventurism emerging from the Islamic Revolution, including the Iranian conflict with

Israel and a new rabid anti-Americanism. The conventional wisdom understands the

Iranian Revolutionary Guards as dispatching to Lebanon in order to “spread the revolution” to the Shia population there as well as to other neighboring and nearby countries, including Afghanistan, Pakistan, Iraq, Bahrain, and Saudi Arabia. This dominant narrative also tells a story of Khomeini’s regime that supported these groups, not only to spread the revolution, but also to “weaken regimes that had banded together to oppose the revolution” (Byman 2005, 37; see also Ramazani 1985, 44). Although this account clearly lines up with those studies that have examined the behavior of post- revolutionary states (see Walt 1989), this story does not account for the one major international event that allowed for the consolidation of the revolutionary regime and shaped the course of Iranian foreign policy and sponsorship in the region at that time: the

Iran-Iraq war.

146 In September 1980, a little more than a year after the Iranian revolution, Iraqi president Saddam Hussein invaded Iran in violation of the Algiers Accord.112 Although he expected it to be a quick victory, this was to become the longest conventional war of the 20th century, costing billions of dollars, millions of lives, and would cease-fire nearly a decade later on exactly the same lines on which it began.

While the massive conventional attack by Saddam Hussein and the inability of the Khomeini regime to easily rearm and resupply created an existential threat to the regime, from the beginning the new Iranian regime began experiencing increased pressure from Iraq. There are indications that the day after the popular referendum on the decision to form an Islamic Republic in Iran, Iraqi aircraft began infiltrating Iranian airspace and later that week on April 7, 1979, reports came in of Iraqi artillery attacks on

Qasr-e-Shirin (Tehrani 1993). Meanwhile, Iran began to train Shia’ Iraqis, especially members of the al Dawa party, and send them back into Iraq where at one point they attempted to assassinate Saddam (Hiro 1991). There is still an academic debate regarding whether the war began because of Saddam Hussein’s desire to take advantage of the weakened Iranian state in order to secure access to the Shatt al’Arab and other geopolitical gains (Johnson 2011; Tehrani 1993) or whether it was in self-defense to force a non-aggression treaty against an increasingly adventurous and ideologically hostile Iran (Al-Marashi & Salama 2008; Aziz 1981). This dissertation does not seek to

112 The Algiers Accord was an agreement reached by Iran and Iraq in 1975 to settle their border disputes, especially over the Shatt-al-Arab waterway. 147 solve this debate, or even really address it. Rather, it takes the long-standing113 and rapidly escalating tension between the two states, and between Iran and the United States and Israel, as the starting point in the examination of Iranian state sponsorship.

While the conflict between Iran and Iraq was clear, conventional, and existential, the conflict with the US was no less of a threat to the new regime. Fearing a counter- revolution supported by the United States, similar to the overthrow of Mohammed

Mosaddegh in 1953 by forces organized by the British M16 and American CIA, and blaming the CIA for the brutal treatment of Khomeini’s followers by the Shah’s SAVAK intelligence service, the regime went on the offensive against the US almost immediately.

While this began as staging anti-US rallies (Randal 1979) and pushing anti-American propaganda, it quickly elevated to the equivalent of diplomatic warfare when a group of students took over the US embassy in 1979--taking 52 Americans hostage--and the regime moved to decisively side with the hostage takers.

Less than a year later, when Saddam Hussein ordered the invasion of Iran, the major powers just stepped back to watch. From the get-go Iran was at a serious supply disadvantage over Iraq. Because of the relationship between the US and Iran during the reign of the Shah, Iran had mostly American made weapons, while Saddam had been reaping the benefits of the regional tilt to the Soviets. Thus, while Iraq could be

113 Iran experienced an HEHI threat environment for 10 out of the 12 years prior to the Iranian Revolution while Iraq faced an HEHI threat environment for 9 out of the 13 prior to Saddam taking the Presidency. While the war brought everything to a head, the conditions already existed for sponsorship to occur. Indeed, both the Shah and al Bakr sponsored terrorist groups. 148 resupplied by any other Soviet client state, Khomeini was reliant on purchasing replacement parts left over from America’s involvement in Vietnam.114

The first several years the major powers stayed out of the fight, yet in 1982, when

Iran pushed the Iraqi army back to the border but refused to negotiate a ceasefire to end the war, the US, France, Great Britain, West Germany, Spain, and many others stepped in to supply a tremendous amount of weapons to the Iraqis, including chemical weapons

(Mesbahi 1993).115 Furthermore, US president Ronald Reagan issued a National Security

Directive (NSD 4-82), which paved the way for the US to begin officially supporting Iraq in the war116 and several other countries followed suit. While the US, like most others, had been largely content to let the two fight it out, fears that the new Islamic Republic could actually win the war sparked an international frenzy of support for Iraq and intensified the material, and existential, crisis for Iran and Khomeini’s revolutionary regime. However, being low on weapons and equipment was not Iran’s only problem, they were also dramatically short on military personnel.

Due to their presumed loyalty to the Shah, most of the officer corps had been killed in purges, forced into early retirement, or imprisoned following the revolution--a serious problem when the war began. Thus, despite the fact that in the early 1980s, Iran held one of the most modern and formidable air forces in the region, there was literally no one to fly the planes. In fact, “Khomeini had to release dozens of pilots from death-

114 Although at the beginning of the war, Iran was also receiving spare parts and weapons through the Israeli government with the knowledge and apparent consent of the Reagan administration. 115 Discussed further below. 116 This was complicated, of course, with the ongoing covert relationship between Iran and the US, including the Iran-Contra affair. 149 row cells, shove instant rehabilitation and pardon certificates into their hands, and beg them to get into the cockpits and win one for the Imam” (Brecher 2003, 2). Even with these “repatriated” pilots, however, it wasn’t long before the Iranian Air Force was virtually destroyed.

Never entirely trusting the loyalty of the released or existing soldiers, early on in the war, Khomeini strengthened his Iranian Revolutionary Guard Corps (IRGC) at their expense. Initially the IRGC was created to ensure the loyalty of those who had served in the military under the Shah and to protect the regime from a counterrevolution (Wehrey et al .2009). While the regular army did approach its pre-1979 strength levels by the mid-

1980s, the IRGC was built in tandem, mirroring, and then quickly surpassing the strength of the regular army.

Figure 5.1: Iranian Regular Army and IRGC force strengths under Khomeini; Source: IISS

150 As the war progressed, the IRGC transformed in both its nature and its mission from a predominately internally-focused and small force, “to the dominant executive agent in Iran’s sponsorship of terror” and transformed Khomeini’s Iran “from a reliance on a strong conventional military to a reliance on terrorism as a means to achieve security” (Prussian 2010, 8). This transition of the IRGC, their mission and their make- up had significant effects on not only the course of the war, but the Iranian military, and the terrorist infrastructure that was built by Khomeini to appease his coalition and maintain his rule.

The war launched by Iraq, however, was not sufficient in itself to transform Iran from a state that relied on a strong conventional army to one that used terrorism in order to achieve security goals, especially that of regime survival. To complete the story, we need to look at the situation and the high level of threat posed to the regime, inside the state.

Nature of the High Internal Threat in Iran The Iranian Revolution and deposing of the Shah had major consequences for

Iran’s regular armed forces. Having been a steady ally, and arms client, of the US for decades as a bulwark against Soviet expansion in the region, the military was suspected of being inherently counterrevolutionary (Byman et al. 2001) and quickly lived up to this reputation. As in the Iraqi and Libyan cases discussed below, there was an established precedent for military coups in Iran that Khomeini had to take into account.117 Iran had

117 Rosemary O’Kane (1981) notes that in states where there have been prior coups, or attempts at coups, future coups are much more likely than if there has been no precedent. 151 experienced two coups prior to the 1979 revolution, both in 1953: one attempt and one that was successful at deposing the democratically elected Mohammed Mossadegh and reinstalling the Shah.118 Indeed, it wasn’t long until “unspecified Army officers” mounted a coup on January 16, 1980, against the new regime and a second attempt was made in

June of that same year (Marshall & Marshall 2010).

Colonel Azar Kahkam organized a third coup attempt on June 27, 1982--not long before the IRGC moved into Lebanon and began to organize the groups that would become Hezbollah (Marshall & Marshall 2010). This attempt followed on the heels of an announcement by the regime that rather than merely push the Iraqi forces back to the border and agree to one of the many cease fire deals being brokered by the UN, Algeria, and others, the Iranian forces were instead “going to liberate Jerusalem, passing through

Karbala” (Sick 1989, 236). This statement caused significant stress and agitation for the more pragmatic factions within the regime, and within the military especially, who were pushing for an end to the war on what would have likely been terms favorable to Iran at that time (Sick 1989).

The refusal to consider an end to the war by Khomeini is widely considered to be a symptom of the radical expansionist agenda of new regime. However, the Theory of

Leadership Survival nudges us towards another explanation, that “as long as the regular army was stretched thin and fully deployed on the western border (with Iraq), it could not mount any sort of a coup d’état against Khomeini and his cohorts” (Wehrey et al. 2009,

24-5). Given these heightened internal threats, the political incentive to consolidate the

118 With the instigation and support of the American CIA and British MI6 152 revolution and secure Khomeini’s rule was a higher priority than military strategy in the war (Wehrey et al. 2009).119 However, the problems balancing the military and consolidating the revolution in the midst of an existential war with Iraq proved tremendously difficult for Khomeini and a fourth coup attempt occurred in May of 1984, once again a result of the frustration of the military leadership over the course of the war.

That these military coups took place in the context of a major conventional war is highly unusual120 and demonstrates the extent to which there was extreme policy divergence between the military and the more radical elements of the regime, one that

Khomeini was forced to mediate. Because he lacked a reliable conventional military that could be strengthened and supported in their fight, Khomeini invested in alternative force structures, such as the IRGC, and terrorist groups that would hit at those who were supplying Iraqi forces in Lebanon or drag them to a second front, and ensured that the future survival of the Iranian revolution was built upon an institutionalized support for terrorism.

Terrorist groups and Leadership survival in Iran The Theory of Leadership Survival argues that terrorist groups are often used by leaders in HEHI threat environments to achieve external security objectives while circumventing the military thus effectively responding to external threats while minimizing the threat the military poses to the regime in the process. The new Iranian

119 Furthermore, Walt (1996) reminds us that in revolutionary regimes, the leader of the coalition is forced constantly to demonstrate his revolutionary credentials and thus the legitimacy of his position in the new regime. 120 Desch (1999) notes that war is the least likely time for the military to attempt a coup, given that the force strength should be reserved for the fight, not for governing the state. 153 regime clearly faced an external threat from Iraq, made significantly worse by the tremendous military and intelligence support from the US, the Soviets, and the West, as well as an internal threat from the regular military. The task now remains to build a case that the regime used their ties with Hezbollah in Lebanon, and groups across the region, to respond directly to these threats, thus protecting Khomeini’s regime and fortifying its position of power, even when those fortifying actions were at the expense of the state itself.

The data seems to indicate that there is an inverse relationship between the number of terrorist groups supported by the Khomeini regime and the number of coup attempts made by the military (see Figure 5.2 below). When Khomeini first took power in 1979 the Shah was already providing support to the Kurdish Democratic Party (KDP) and the Palestine Liberation Organization (PLO) (San Akca 2009; Salahyan et al. 2009) a trend that Khomeini continued. After the revolution, Khomeini faced two coup attempts in 1980 and by 1981 Iran was providing support to an additional group. That level of support held steady until 1982 when there was yet another coup attempt. In 1983, the number of supported groups increased to four, including Hezbollah in Lebanon, and by the fourth coup attempt in 1984, Khomeini was supporting five terrorist groups. After the last coup attempt in 1984, the number continued to increase.

154

Figure 5.2: Number of terrorist groups supported and coup attempts faced by Khomeini, by year (1979-1989)

Sources: Marshall & Marshall (2010) and the author.

Although Khomeini supported a number of diverse terrorist groups during the war

—that were actively perpetrating terrorist attacks at the time--there are three sponsored groups that are worth examining here in relation to securing Khomeini’s position of power: first, the PLO, second, the anti-Iraq Kurdish groups and finally, Hezbollah. While most scholars focus on relations between Iran and other Islamic groups in the region, and world wide, there are indications that even before the creation of Hezbollah, the new

Iranian regime had strong ties to Lebanon. Sick (1991) argues that the Palestine

Liberation Organization (PLO) had, in fact, been involved with and “had trained many of the Iranian revolutionaries in Lebanon in the years prior to the revolution” and also notes that “PLO operatives provided communications, security, and other technical services

155 and support for the new regime” (74-5). Yasser Arafat himself, in fact, reportedly boasted of having helped to train and organize the IRGC (Cann & Danopoulos 1998, 276).121

Given these pre-existing links, it is not surprising that Iranian support to the PLO continued once Khomeini was in power.

Figure 5.3: Active terrorist groups supported by Khomeini (1979-1989)

It is, however, a bit of a puzzle that Khomeini’s regime, desperately trying to consolidate power following a social revolution and tremendous domestic turmoil, while facing both an existential threat from Iraq--in the form of conventional war that the

Iranian state was not at all equipped or prepared to fight—as well as a tremendous internal threat from the very same military the regime required in order to repel that threat, would dispatch valuable IRGC cadres to, and invest resources in, Lebanon—a

121 In addition, one of the key figures of Hezbollah and mastermind of many of its terrorist operations, Imad Mughniyeh, was active in Fatah’s Force 17 before Hezbollah was fully organized (Ranstorp 1997). 156 state not even bordering Iraq. Perhaps the most important dynamic to consider as we take a closer look at the Theory of Leadership Survival is Iran’s relationship with Hezbollah.

Prior to organizing Hezbollah, Iran was already supporting groups in Lebanon, not just the PLO, but also providing support for the Shia Amal (Miller & Smarick 2012) and some other local groups including the Islamic Dawa Party, one of the first terrorist groups to target Iraq (Claiborne 1980). Although there were indications that Iran was arming and supporting the folks that would become Hezbollah by 1980 (SIPRI); it was following the 1982 push-back of Iraqi troops to the border, the third coup attempt in Iran, and the NSD 4-82 authorizing US support to Iraq—all of which coincided with the Israeli invasion of Lebanon--that activity really picked up. By 1983, there were some 650 IRGC personnel in Lebanon (IISS 1983, 55).

Several of the most famous Hezbollah-linked attacks took place during this time: the Israeli embassy bombing, the US embassy bombings and the Marine barracks bombing.122 The embassy housed not only diplomatic staff, but also housed the CIA, the office largely responsible for the arms transfers to Iraq during the war.123 While it is likely that Khomeini seized the opportunity to “export the revolution” to the Shia population in Lebanon and show that they were capable of taking the war to Jerusalem, it is also widely accepted that these attacks were perpetrated to drive the West (and Israel)

122 The Iraqi embassy was also bombed in Beirut in 1981. While it is still unknown who exactly perpetrated the attack, there were indications at the time that it was either Iraqi Kurdish groups or members of the al- Dawa party, both of which were supported by Iran. 123 “The CIA, including both CIA Director Casey and Deputy DirectorGates, knew of, approved of, and assisted in the sale of non-U.S. originmilitary weapons, ammunition and vehicles to Iraq.” Former NSC official Howard Teicher, testimony in US v Cardoen, 1995.

157 out of Lebanon. There are also indications that Hezbollah’s terrorist attacks explicitly targeted those countries that were supplying Iraq with money and weapons against Iran in the war, a clear military objective that was being achieved without the use of a military that was 1) already engaged in a long-term conventional war with Iraq; and 2) unreliable in terms of trying to seize domestic political power from the regime.

Between 1982 and 1988, the Global Terrorism Database (GTD) records 87 attacks perpetrated by Hezbollah inside Lebanon and around the world, while other sources, such as Pape (2005), Ranstrop (1997) and the Nations Encyclopedia project124 provide data on

51 more (see Figure 5.4 below). Participation in the Lebanese civil war (not counting attacks perpetrated against the Israeli proxy South Lebanese Army)125 accounted for 13

(or about 9.5%) of these attacks, while 16 attacks (11.7%) were against Israeli military convoys and bases in Lebanon, and Jewish civilian targets elsewhere. A full 12.4% (17 attacks) were perpetrated against the Christian South Lebanon Army, an Israeli proxy.

Yet, the overwhelming number of attacks (78 or 56.9%) explicitly targeted countries that were directly providing military, financial, or logistical support to Iraq.

France alone was responsible for providing Iraq with high-tech arms, including

Mirage F-1 fighters, valued at anywhere between $5.1 billion (Metz 1988) to $5.6 billion

(Mesbahi 1993) between 1981 and 1985, which constituted approximately 40% of all

French arms exports (Metz 1988). In addition, France (target of 15 Hezbollah attacks,

124 http://www.country-data.com/cgi-bin/query/r-8105.html 125 For the sake of a conservative assessment of Hezbollah’s role in Iranian foreign policy during the war, the SLA is assumed to be acting in the interest of Israel and not as a faction within Lebanon’s civil war. 158 10.9% of total) and Italy (target of 1 attack by Hezbollah) provided Saddam Hussein assistance with the construction of the Osirak nuclear facility (Johnson 2011).

While the Soviet Union was responsible for the bulk of the military aid to Iraq

(Mesbahi 1993),126 the KGB was rather successful in deterring Hezbollah attacks. When four Soviet embassy officials were seized (the two attacks noted in Figure 5.4 below) and one killed, the Soviets responded by having the KGB compile a list of blood relatives of the Hezbollah leadership. The KBG then proceeded to kidnap one, castrate him, and send his severed genitals back to Hezbollah with a note listing the names and addresses of their other relatives and stating unless the remaining three attaches were released, the trend would continue (Mckinney 1988).127

126 Including more than 2,000 tanks, 300 fighter aircraft, nearly 300 SCUD missiles and thousands of heavy artillery rounds. 127 The remaining three officials were dropped off at the Soviet Embassy in Beirut immediately following this exchange. 159 Hezbollah Attacks (1982-1988) by Target State

35 30 25 20 15 10 5 0

Figure 5.4: Hezbollah Attacks (1982-1988) by Target State.

The United States, which was the target of 30 Hezbollah attacks between 1982 and 1988 (21.9% of all attacks), was not only responsible for providing the Iraqis with billions of dollars of credits, U.S. military intelligence and advice (Teicher 1995), but also provided Saddam Hussein with chemical weapons (Mesbahi 1993). The 378 chemical weapons attacks128 occurred not only on the battlefield, but also against Iranian civilians during the war of the cities. Between 1983 and 1988 these attacks resulted in

5,793 dead and 43,973 wounded (Shemirani 1993). Iran, specifically the IRGC in

Lebanon that had direct ties to the clerical establishment and convergence of interest with

Hezbollah (Ranstrop 1997), had clear incentives to stop the supply of US assistance to

Iraq.

128 Including nerve gas, blister gas, hydrogen cyanide, and blood affecting gasses. 160 Hezbollah attacks between 1983 and 1988 targeted US embassies, the regional

CIA office, US Marines, and resulted in the kidnapping of multiple Americans—many of which were later released in the infamous “arms for hostages” deals made by the Reagan administration. Indeed, rather than being a rhetorical tool, the assertion that the war was fundamentally orchestrated and sponsored by the United States was a widely held certitude among key decision-making circles within the regime (Mesbahi 1993). If

Hezbollah was designed primarily to attack Israel, then 66.9% of the time during the Iran-

Iraq war, it was not on task.

Examinations of the other targets reveal a similar pattern. Kuwait (target of 9 attacks, or 6.6%, by Hezbollah) and Saudi Arabia (target of 6 attacks, or 4.4% of all attacks between 1982-1988) largely bankrolled Iraq, with estimates of loans and grants topping $20 billion a piece. Both states were also targeted by the Iranian naval forces during the Tanker War, prompting not only the US reflagging of ships and increased US naval protection, but also the release of the now declassified NSDD 141 in 1984 that called for the US to increase arms (including STINGER missiles) and tanker support to

Saudi Arabia. Great Britain (target of 8 attacks, 5.8%) likewise provided Iraq with weapons and Jordan (target of 2 Hezbollah attacks) provided logistical support for Soviet arms being transported through the Jordanian port of Aqaba (Mesbahi 1993).129

129 The four attacks against Syria all occurred after the 1985 crackdowns, including a threat by the Syrian regime to kill Sheikh Fadlallah unless they stopped their overt attempts to establish Lebanon as an Islamic Republic. From then on out, despite the Syrian-Iranian alliance, the relationship between Hezbollah and Syria became “characterized by periods of conflict and cooperation” (Ranstorp 1997, 71). 161 That vital support for Iraq was being provided by the Soviet Union, the conservative Monarchies and the West was well understood by Iran. The Islamic

Republic of Iran Plan and Budget Organization released a study of the economic damages of what Iran called the “Imposed War” in 1983. Contained within the document is a passage that shows clear recognition that Saddam’s survival is fundamentally dependent upon this aid:

And Saddam who was to deliver his victory speech in Ahwaz on August 25, 1980 (3rd of Shahrivar 1359) is now forced to beg and beseech one or the other of the pro-American Sheikhs in the Persian Gulf on the one hand and the United States, the Soviet Union and France on the other hand in order to protect his rule for a few more days (Islamic Republic of Iran Plan and Budget Organization 1983, 12).

With the exception of thirteen attacks on the Israeli military in Lebanon, seventeen on the

South Lebanese Army,130 and three more on Jewish targets internationally, nearly every single one of the attacks perpetrated by Hezbollah between 1982 and 1988 targeted a state that provided military, financial, and/or logistical support to Iraq.

While the groups that Khomeini, via the newly empowered IRGC, provided support to in Lebanon, including Hezbollah worked to punish countries that provided Iraq with support and sever these supply lines, the Kurdish Iranian-sponsored groups--The

Kurdish Democratic Party (KDP), the Patriotic Union of Kurdistan (PUK), and the

Kurdish Workers Party (PKK)--opened up a second front for the Iraqi army in the north.

130 Which was a Lebanese Christian organization that was funded, armed, and trained by the Israeli government, but who also fought Hezbollah directly for the control of southern Lebanon. While this study assumes attacks on the SLA were “proxy attacks” on Israel, so as not to overinflate the participation by Hezbollah in the Lebanese civil war and potentially under-examine the prevalence of Israel as a target of these attacks. 162 The Kurdish Democratic Party was receiving money, weapons, and logistical support from the Shah as early as 1968 (San Akca 2009), who was using their attacks to gain leverage with the Iraqis, and that support did not cease when Khomeini took power.

By 1986, the PUK had joined with other Kurdish forces in the Iraqi north and were waging a battle against the Iraqi regime that split their army forcing them to fight on two fronts (al-Marashi & Salama 2008) and once again helping to minimize Khomeini’s reliance on Iran’s regular forces to fight the war.

Indeed, the KDP was such an asset in the war by 1987 that the IISS Military

Balance inventory lists the 12,000-man force as a paramilitary unit of Iran (IISS 1987).

Although the Global Terrorism Database (GTD) shows only four attacks during the course of the Iran-Iraq war131 this is likely an indication of holes in the dataset132 rather than low frequency of KDP attacks. Because the GTD is used to set the criteria in this study for what constitutes a terrorist group, however, the fact that the official attack numbers are low and proportioned toward military attacks prevents these Kurdish groups from being officially used in support of the theory. Even if the GTD numbers are correct, or biased toward military attacks generally, it may indicate that this theory can stretch beyond terrorist groups to demonstrate the use of non-state armed groups as a tool of regime survival.

131 Two against military targets exclusively; one against military/private persons and property; and one against a government target. See http://www.start.umd.edu/gtd/search/Results.aspx?perpetrator=699 accessed July 9, 2012. 132 As discussed in Chapter 4. 163 The Iranian case, then, seems to generally confirm what the Theory of Leadership

Survival would predict, that facing severe and concurrent external and internal threats, the leader would build up alternative forces to the regular military, especially intelligence services, and would use these terrorist groups that remained outside the control of the military in order to achieve policy goals that were vital to demonstrate the leader’s competence to the other members of the ruling coalition, strike at the external enemy, and decrease the threat posed to the ruler by the regular military by keeping it weak.

Ironically, as also seems to be the case with Pakistan, the very organization that was empowered to enact these policies to help shore up the regime, the IRGC, was also the organization that over time consolidated this power and went on to pose the greatest threat to the regime.133

SADDAM HUSSEIN’S IRAQ Unlike the revolutionary beginnings of Iran’s Ayatollah Khomeini, Saddam

Hussein rose to power within the ranks of the Ba’ath party establishment in Iraq to take control of the Revolutionary Command Council (RCC). Known for his ruthlessness and shrewd political skills, once Saddam took the Presidency, he was able to consolidate power rather quickly, in part because he had already been working behind the scenes to establish alternative intelligence networks that could not only “check the General

Security Service, but…counter balance the Military Intelligence as well” (al-Marashi &

133 This dynamic of security services and other branch organizations entrenching their power vis-à-vis the civilian leadership through sponsoring terrorist groups is a dynamic that is beyond the bounds of this study, however, Chapter 7 examines ways to explore this more fully in future works and argues that understanding this dynamic is fundamental to gaining a complete understanding of state sponsorship. 164 Salama 2008, 120). He had been building up the coup-proofing infrastructure of Abu

Bakr’s regime, only to one day seize it himself and the violent purge of the RCC in 1979 immediately following his ascension,134 indicated that he was more than a little aware of the holes in the system.

Saddam was no stranger to the harshness of Iraqi politics. Politically active at an early age, he took part in a failed assassination attempt of the Iraqi Prime Minister in

1959 was forced to flee to Syria. The assassination attempt marked the beginning of a long relationship Saddam had with the complexities of staying in power, first from the standpoint of trying to depose an existing ruler, and later, struggling with the issue of how to maintain that position of power for himself. Saddam consolidated power within the Ba’ath party so well, that when Abu Bakr stepped down in July of 1979, few Iraqi’s believed the official story of Bakr retiring for health reasons, but rather believed Saddam had staged a coup within the Ba’ath party after receiving assurances from the British and

Americans that they would not interfere (Aburish 2000).

Although Saddam was a new leader coming to power in an HEHI threat environment,135 one cannot completely understand Iraqi state sponsorship by starting with

Saddam Hussein. By the time Saddam took power, he inherited a sponsorship infrastructure developed by Abu Bakr who was already supporting a number of groups including Abu Nidal (Adams 1986; Melman 2003; UCDP External support dataset; Seale

1992), the Palestinian Front for the Liberation of Palestine (Andrew & Moitrokhin 2005;

134 That resulted in the immediate execution of five RCC members and 17 other of Saddam’s rivals he accused of being involved in a Syrian plot to stage a coup against his new regime. 135 As will explained in the sections below. 165 UCDP External support dataset), the Irish Republican Army, the Basque separatist group

ETA, the German Red Army Faction, the Japanese Red Army, Tupamaros (Ginat & Bar

Noi 2007), and the Palestine Liberation Organization (Fontaine 1988; Seale 1992;

Sterling 1981). Like Saddam Hussein, Bakr also faced a HEHI threat environment: one coup attempt in 1973 and a number of Militarized International Disputes (MIDs) during his rule (1969-1979). When Saddam took control of the RCC and the Iraqi state in 1979, he walked into an already institutionalized system of state sponsorship to secure and fortify the leadership, much of which was of his predecessor’s making.

Threat Environment for Saddam Hussein Saddam Hussein was in power in Iraq for 25 years (1979-2003) and although he faced an HEHI threat environment in general, there were three periods of distinct HEHI threat, two of which will be examined here. The first period took place during the Iran-

Iraq war. The second major HEHI threat period followed Saddam’s 1990 invasion of

Kuwait and the international coalition’s mobilization to force his withdrawal. While the third period of HEHI threat followed the coup plot by Bridiger General Abd al-Karim al-

Dulaymi in January of 2000 and stretches into the run-up to the second Iraq war. The first two threat periods will be the primary focus of the analysis in this chapter. Due to lack of existing threat environment data that stretches into the mid-2000s, the third period of threat will be examined in-depth in future work.

166 Iran-Iraq War and External Threat to Saddam’s Regime When Iraq sent the first contingent of 70,000 troops across the Iranian border in

September of 1980, it was the largest deployment of Iraqi troops in entirety of the war.136

Al-Marashi & Salama (2008) argue that Saddam Hussein’s objectives in invading Iran were three-fold: first, counter the threat to his regime emanating from revolutionary Iran, which Aburish (2000) argues was “so real that there was little room for a modus vivendi between the two countries” (192); second, take full control over the disputed Shatt Al-

Arab; and finally, to demonstrate Iraq’s military prowess. Regardless of the initial reasons for the invasion, the quick victory that Saddam expected did not materialize and he had to rapidly build and equip a much larger conventional force in order to repel the

Iranian counter-invasion.

The initial investment in the war was so tremendous, and reliable military aid so spotty, that within the first couple of years Iraq became destitute. When in 1982, Saddam tried to call for a cease-fire and Khomeini refused--instead taking the offense and attempting to seize the Shi’ite dominated city of Basra, a number of countries stepped in to provide Iraq with weapons, equipment, intelligence, and training. Although the attempt to seize Basra failed, the lines solidified as both sides dug in, literally, holding each other at bay while fiercely attempting to break the stalemate for the next several years.

136 Again, whether this was intended by Saddam to send a message to the Ayatollah Khomeini that Iraq would not be intimidated by his rhetoric that called for the Iraqi President to be overthrown (al-Marashi & Salama 2008), a land-grab by an expansionist and ambitious leader taking advantage of Iranian weakness (Rajaee 1993) or a combination of the two will not be debated here. 167

Figure 5.5: Iraqi Regular Army Personnel (1979-1989); Source: IISS 1979-1989

After the Iranian counter-invasion of 1982, concerned that the Iranians could actually win the war and upset the regional balance, the Reagan administration removed

Iraq from the list of state sponsors of terrorism. Reagan also authorized a National

Security Decision Directive (NSDD 4-82), an executive order that permitted the US to sell weapons to Iraq. In short order, the US, Soviet Union, France, and various other

European and Arab countries stepped in to prop up Saddam and helped him rapidly double the size of his army. Indeed, as Figure 5.5 indicates above, the regular armed forces went from approximately 300,000 in 1982 to over 600,000 personnel in 1984 and by the time the war ended in 1988, there were more than 1 million Iraqi’s under arms (a

500% increase from 1980).

168 This rapid increase in the size of the Iraqi army also increased the threat that the military posed to Saddam’s regime and, in fact, the army mounted a coup against him on

December 20th, 1984. Although Saddam survived, an estimated 62 people were killed in the coup attempt (Marshall & Marshall 2010). In response, Saddam diverted significant resources towards a massive parallel build-up of the elite anti-coup force, the Republican

Guard, and Saddam began increasingly to rely on terrorist groups to achieve political objectives over the course of the war.

Internal threat and the Iran-Iraq War The war against Iran was also partially intended to entrench Saddam’s power within the RCC and came on the heels of a coup attempt by several military officers including A. Hussein al-Hamdani, M.A.H. Mashhadi, M. Mahjoub, M. Ayesh, and G.

Abdel-Jalil, just twelve days after his ascent to the Presidency in July of 1979 (Marshall

& Marshall 2010). Iraq, like their Syrian Ba’athist neighbors, had faced wave after wave of military coups. The threat of coup, in fact, played a large role in the institutional structure of the Ba’athist state itself and the RCC. The Ba’athists that came to power in

1968, Saddam among them, were very “aware that the military (had) been the primary force behind every regime change in Iraq since 1936” (al-Marashi and Salama 2008,

112). Indeed, between 1958, when Iraq gained its independence, and the time the Saddam took full power of the state in 1979, there had been eight coup attempts (five in 1965 alone) and four successful coups d’état in Iraq (Marshall & Marshall 2010).

169

Figure 5.6: Coups in Iraq prior to Saddam assuming the Presidency (1958-1978)

When Saddam took the Presidency, the state had in the process of shifting from military rule to Ba’ath party rule with an increasingly politicized army that was being forced into subordination under the RCC through a series of purges, officer rotations, and appointments to high office of family, party, and clan loyalists. Still, it was clear that the party’s main “support within the state essentially emanated from the military” (al-

Marashi and Salama 2008, 94), thus making the military also the best suited to undermine the regime. One of Saddam’s first actions once assuming the President’s office was to increase the salaries of the military and to ensure that “the generals and officers who had survived the purges owed their positions to loyalty to Hussein rather than to any military competence” (al-Marashi and Salama 2008, 127). While this appeasement (buying off) of the upper echelons of the military and other motions to ensure their loyalty took care of

170 the senior officers, these moves did not adequately address the threats from ambitious junior officers.

Even with these “safeguards” in place, the threat from the military continued and once in power, Saddam faced six137 additional plots and coup attempts to overthrow his rule (Marshall & Marshall 2010). Indeed Saddam rose thought the ranks in what could only be considered a culture of extreme coups; so much so that there was a long-standing

“Iraqi tradition of preventing military units’ access to ammunition and fuel” with the exception of the Revolutionary Guards (RG) who functioned outside the military chain of command and were intended originally to serve in opposition to the military as an anti- coup force (al-Marashi and Salama 2008, 120). By the time of the Iran-Iraq war,

Saddam was walking a very delicate, and increasingly difficult, line between providing the army with enough weapons, training, and manpower to successfully fight the war, and preventing them from marching into Baghdad.

The role of the Republican Guard Force in Saddam’s regime survival and in response to resolving this internal threat cannot be over-stated. Although it started out as a rather small force, only one mechanized brigade in 1979 when Saddam took power

(IISS 1979), following the 1984 coup attempt, Saddam diverted tremendous resources to the RG, a force that could keep the regular army “in check” (al Marashi & Salama 2008,

167). By 1985, it stood at two armored brigades, one infantry brigade, and a commando brigade rivaling the size of the regular forces (IISS 1985).

137 Although only one more during the Iran-Iraq war, in 1984. 171

Figure 5.7: Strength of the Republican Guard (in brigades) and Coup attempts against Saddam Hussein (1979-1989)

Source: Coup data is from Marshall & Marshall (2010); the Republican Guard data is from the IISS (1978/79 through 1988/89).

Given the tremendously high internal and external threat environment that

Saddam faced during the Iran-Iraq war, a similar puzzle to Iran’s regarding Iraq’s support to terrorism arises: Iraq is forced into massive mobilization in an existential international war that Iraq was not clearly winning and could potentially become bankrupt losing, why did Saddam invest resources in groups that were operating “out of theatre” in places like

Israel and Lebanon? Alternatively, once the great powers intervened and there was such a tremendous increase in Iraq’s military power, why would Saddam need to employ terrorists?

172 Terrorist Group support and Regime Survival during the Iran-Iraq War One of the best anti-coup measures a leader can take is to develop “anti-armies” that report directly to the executive, circumventing the military command, and function as a check on military power (Quinlivan 1999). This study argues that the terrorist groups operating with the support of Iraqi intelligence services functioned as (and bolstered) additional “anti-armies” that put some of Saddam Hussein’s capabilities to achieve vital policy objective outside of the control of the military, allowing Saddam to both prevent his loss in the war and off-set some of the increased threat posed by increasing resources to the Iraqi army.

During the early years of the war, Saddam provided support to groups such as

Abu Nidal,138 the Palestine Liberation Organization (PLO), the Arab Liberation Front

(ALF), and the Syrian Muslim Brotherhood. In the very early stages of the war (1981-

1982), the Muslim Brotherhood launched a series of attacks in Syria, Iran’s only ally in the region, and mounted an insurgency that resulted in the obliteration of the Syrian city of Hama, all while decreasing the ability of Syria to provide aid to Iran’s war effort.139

However, it was the support to the Mujahedeen-e-Kalq that was the most vital to the war and to Saddam’s regime survival.

138 As part of the agreement with the US to remove Iraq from the state sponsors of terrorism list, Saddam agreed to cease support for Abu Nidal and even had him expelled from Iraq. He relocated temporarily to Libya, but by the 1990s had returned to Baghdad. 139 Syrian leader Hafez al ‘Assad was also more oriented towards Lebanon at the time and the conflict being waged there between the PLO and the Israeli government. In any case, Syria was kept seriously constrained in their ability to assist Iran, in part because of internal turmoil caused by the Muslim Brotherhood uprising. 173

Figure 5.8: Number of attacks perpetrated by select groups Saddam was supporting during the Iran-Iraq war.

By 1983-84, following the tremendous build-up of the Iraqi army,140 Saddam began to show “quirky military behavior” by providing significant support to the MeK, even to the point of relocating them to bases within Iraq (Aburish 2000, 231). While

Saddam Hussein began funding the MeK in 1980, by 1986 he was providing the MeK with “protection, funding, weapons, ammunition, vehicles, tanks, military training, and the use (but not ownership) of land” (Goulka et al. 2009, 3).

140 Keep in mind that the Iraqi army doubled in size during this period, from 300,000 in 1982 to 600,000 in 1984. 174

Figure 5.9: Support given the MeK by Saddam Hussein during the Iran-Iraq war (1979- 1989)

Many Iraqi generals, already subject to extreme duty rotations and given very little operational freedoms, were dismayed at the resources being diverted to the MeK and had a general feeling that, “Saddam made more of this (relationship with the MeK) than was militarily justifiable” (Aburish 2000, 231). The MeK crossed over the border into Iran numerous times to fight Iranian military forces and the IRGC (Goulka et al.

2009) as well as perpetrated terrorist attacks inside Iran. Only about 12% of MeK attacks during the war targeted the Iranian military forces, while another approximately 41% of attacks were focused on government targets, yet the bulk (47%) of the attacks were on non-military/non-government targets such as businesses, utilities, and private citizens and property (GTD).

175

Figure 5.10: Target patterns by the MeK inside Iran during the Iran-Iraq war;

Source: START Global Terrorism Database

The conventional wisdom holds that states sponsor terrorist groups when they’re weak, yet this is an instance of an increase in support to the MeK when the Iraqi army was stronger than it had ever been. Why would Iraq, with the fourth largest military in the world at the time, decide to fund and dispatch and Iranian dissident group to attack

Iranian armed forces and perpetrate terrorist attacks inside Iran (Goulka et al. 2009) rather than provide greater operational freedom, weapons, and intelligence to one of the world’s largest conventional armies?

The traditional explanations of a state fighting with a conventionally weak military or of a leader relying on the “deniability” of their actions to pursue policy objectives do not fully explain the sponsorship patterns of Saddam Hussein. This section

176 has provided an alternative account of a situation where Saddam could not trust the military that he had to mobilize to fight an external threat. In addition to strengthening the

Republican Guards as a parallel army, he chose to invest some of Iraq’s capabilities in the

Iraqi intelligence services and terrorist groups they armed and trained to strike at Iran and its allies, especially the MeK. This not only kept some of Saddam’s ability to meet foreign policy objectives outside of military control, but also helped him to prove himself a strong and capable leader within the RCC and weakened Iraq’s adversaries while balancing the internal threat from the military. This pattern, as the next section shows, continued into ensuring the survival of his regime after the Gulf War.

Operation Desert Shield/Desert Storm and Saddam’s External Threat Following the end of the Iran-Iraq war, Saddam presided over a weak and stumbling state—with high employment and food shortages (Aburish 2000), low oil prices, the war debt past the ceiling, rampant corruption, the economy bankrupt, and a battle-hardened army of nearly one million men with literally nowhere to go. Believing that he led Iraq to singlehandedly save the Gulf states from an Iranian onslaught, Saddam demanded that all war debts be erased, something that the monarchies refused (Khalidi

1991). Seizing once again on the long-standing irredentist claim that Kuwait was really part of the Basra province of Iraq,141 Saddam calculated that taking Kuwait would immediately erase roughly US$20 billion dollars of his standing Iran-Iraq war debt

(Aburish 2000; Khalidi 1991), provide access to the gulf for oil exports (Frankel 1991),

141 Kuwait was carved off from of the Basra province by the British in 1899 and when Iraq was granted independence in 1922, Kuwait was intentionally not included in the new state because the access to the gulf that it would afford Iraq was considered a challenge to British dominance in the region (Frankel 1991). 177 and ease the pressure from the Iraqi military that was now bursting at the seams (Aburish

2000; al-Marashi & Salama 2008).142

While it was Saddam that made the choice to invade Kuwait and thereby increase the level of external threat,143 given the deteriorating domestic situation, it was a move to protect his rule, help relieve some of the debt burden from the war and release some of the pressures of a massive military that was tremendously unsatisfied with the regime’s ability to meet their high post-war expectations regarding the “fruits of victory” (Aburish

2000, 253), could not be demobilized, and threatened to turn in on the regime (al-Marashi

& Salama 2008, 176).144

In perhaps simply another series of miscalculations, Saddam misread the likely

Western response and failed to take and hold Kuwait without objection. It wasn’t long before UN Security Council Resolution 678 was passed and an international coalition, led by the United States, began to amass an army in Saudi Arabia to push him out of Kuwait.

The war itself was over in record time, 100 hours, and left Saddam penned in by no-fly zones, sanctions, constant monitoring, and occasional airstrikes by the US and Great

Britain. Saddam was now in a position of significant external threat, especially from opportunistic neighbors, and a growing internal threat from a smaller, but greatly disgruntled military.

142 Although some argue that the seizure of Kuwait was not Saddam’s ultimate goal, but rather Saudi Arabia (Krauthammer 1991). 143 Indeed, there are some indications that leaders who face an acute internal threat are more likely to start external wars. 144 The precise relationship between internal and external threats is also something that should be explored in future studies. 178 Internal Threat and the Gulf War The influx of massive military aid to bolster the regime during the Iran-Iraq War had allowed Saddam to build an army that was more highly trained and better armed than ever before. During the war, he had increased the size of the Republican Guard (and supported terrorist groups) in a way that kept pace with the rapid growth of the army, and now with the collapse of oil prices and rampant unemployment, the demobilization of the army was a virtually impossible task. When Saddam gingerly attempted the demobilization of around 200,000 soldiers, the program had to be quickly aborted because “discharged men, desperate for jobs started street brawls with Egyptian workers in which dozens died;” there simply were no civilian jobs to absorb them (Aburish 2000,

253 & 261). While the Gulf war reduced the Iraqi regular army almost to its pre-Iran-Iraq war numbers, officers were even more dissatisfied with Saddam’s rule.

Iraqi Regular Army Personnel (1979-2003)

1200 1000 800 600 400 In thousands 200 0

Figure 5.11: Number of Iraqi Regular Army Personnel, in thousands (1979-2003).

179

After the invasion of Kuwait and the devastating loss on the battlefield against the

US coalition forces, Saddam faced three additional coup attempts, two in 1991 and a third in 1992 (Marshall & Marshall 2010). When Saddam continued to make counter moves to strengthen the Republican Guard at the expense of the regular military, some of these threats began emanating from the Republican Guard itself. In response, he developed yet another “parallel force” that reported directly to him, once again bypassing the Ministry of Defense (al-Marashi & Salama 2008, 187). With the external threat high and the internal threat reaching tremendous levels, Saddam once again turned to terrorist groups to help fortify his rule.

Terrorist group support and leadership survival in the Gulf War Following the Gulf War, Saddam needed to solve three overlapping problems in order to secure his rule. First, he had to regain control over the country, reconsolidate his rule and suppress the insurrections in the Kurdish North and the Shia South. Second, he needed to prevent the military from mounting additional coups and the threats arising from within coup-proofing forces of the Republican Guard were especially concerning.

Finally, he needed to minimize the external threat to his regime from opportunist neighbors, especially Iran, and demonstrate that although the Americans had pushed him out of Kuwait he was still a regional power. He used terrorist groups to help achieve all three.

Between 1990 and 1992, the years of the most acute external and internal crises, attacks by groups supported by Saddam Hussein skyrocketed, especially attacks by the 180 MeK against Iran and the PKK against Turkey. Again, this poses a puzzle, why would a leader who faces an extreme HEHI threat, was boxed in by an international coalition of some of the strongest armies on earth and faced three coup attempts in 1991 & 1992 decide to invest finite, and quickly dwindling, resources in terrorist groups?

Figure 5.12: Attacks perpetrated in Iran by the MeK (1990-2002)

Note: Saddam faced two coups in 1991; one in 1992; and one in 2000.

I argue that these groups allowed Saddam to meet all three of his vital objectives to remain in power. First, they assisted with putting down the insurrection and helped him to reconsolidate his regime, second, they helped him weaken his main external threats and demonstrate strength, and finally, they allowed him to do this without depending on the goodwill of the restless junior officers who were demonstratedly unreliable, as they were now mounting coups against his rule.

181 The role that terrorist groups played in putting down the insurrections in Iraq lies outside the bounds of this theory, but is worth examining nonetheless. By the definition developed in Chapter 2 and used in this study, terrorist groups are those that act outside of the borders of the state, however, in addition to the cross-border attacks perpetrated against the Iranian forces and population during the Iran-Iraq war, and the spike in attacks they perpetrated against Iranian government, military, and civilian targets after the war, there are indications that the MeK also participated in putting down the rebellions—indeed one source states that 15 of 18 provinces rebelled following the pull- back from Kuwait (al-Marashi and Salama 2008, 184)—in the Kurdish North and Shia south after the 1991 Gulf War (Goulka et al. 2009). Saddam’s distrust of the military in general during this time is illustrated by the use of the security services and Republican

Guard to put down this intifada and the fact that it was “the inner core of his family and the party faithful” that handled the operations by the regular Iraqi army to put down the insurrections (Aburish 2000, 310).145 The MeK provided Saddam with another force to deal with this internal (partially military) threat to his rule.146

The external threat for Saddam was also particularly acute in this time period. In fact, unlike during other international crises, Saddam himself was the target of the sanctions with the Western powers being very clear that they would drop the sanctions and end the no-fly zones only once Saddam was no longer in power. The northern no-fly

145 Additionally, Aburish (2000) argues that many of the “rebels” were actually defected soldiers to whom Saddam offered amnesty if they would rejoin their units. Following the suppression of the intifada the officers that had participated were executed without trial. 146 While the Theory of Leadership Survival does not fully account for the effects of internal war on state sponsorship, the quantitative analysis suggests that it is a statistically significant factor in a leader’s decision to sponsor. 182 zone covered the Kurdish areas in the north and patrols and attacks were largely staged in

Turkey (Ricks 2000), while the southern no-fly zone covered the Shi’ite areas of the Iraqi south and was managed from the Ali Al Salem and Al Jaber Air Bases in Kuwait. In 1991 and 1992 especially, Saddam struggled to minimize the external threat posed by his neighbors—who could prey on his weakness—and to neutralize the effects that the no-fly zones and sanctions regimes enforced by the British and the Americans were having on his rule. As in the period between the end of the Iran-Iraq war and the invasion of

Kuwait, Saddam turned back towards rabidly demonizing Israel, providing support for the PLO147 and other Palestinian groups in order to demonstrate to the RCC that he was a capable leader and that Iraq was still a regional power.

The third threat that he faced to his rule during this time was from within the military. He faced two coup attempts in 1991 and an ambush on his motorcade in 1992

(Marshall & Marshall 2010). Not only did Saddam make further moves to install family members in key positions in the military and security services, but he carved off a slice of the Republican Guard in 1992, called the Golden Division, that was “paid higher salaries and accorded priority over normal mortals even in getting food and prescription drugs”

(Aburish 2000, 325). Despite these “coup-proofing” measures, the critical need to omnibalance the multiple origins of threat to Saddam’s rule once again led him to support terrorism. The Theory of Leadership Survival predicts that when the HEHI threat is high and omnibalancing required, support for terrorist groups will increase (along with the

147 To whom Saddam gave “more money than he could afford” (Aburish 2000, 254). 183 buildup of other parallel forces and even unconventional weapons). Indeed, during this time period, attacks by Iraqi-supported groups began to skyrocket.

The MeK was not the only group receiving Iraqi support. Saddam also resumed ties with Abu Nidal, continued to aid Palestinian groups such as the PFLP and the PLO and began to support the Kurdish PKK. Specifically, when Turkey chose to side with the international coalition to force Saddam out of Kuwait, he began to arm the PKK (Canci &

Sen 2010). There are indications that the massive upsurge in PKK attacks in 1992 was a combination of Turkish participation as a staging ground for the US no-fly zone established in the north of Iraq and the weapons made available to the PKK by Saddam

Hussein and that this support was an attempt to check the external threat to his rule posed by Turkey and the US.

Figure 5.13: PKK targets and number of attacks while receiving support from Saddam Hussein (1990-1995) 184 Again, while it is difficult to draw the line directly from provided support to perpetrated attacks, there are indications that the scale, and perhaps number, of attacks are signals of terrorist resources (Overgaard 1994), many of which come from national governments. There are also indications that the length of time between clusters of attacks relate to not only counter-terrorism operations, but also the need to renew these resources (Sandler & Enders 2003). Interestingly, it was not only the MeK and PKK that had high levels of attacks during the years that Saddam struggled to reconsolidate his rule. Indeed, the PFLP and Abu Nidal also show spikes during 1992.

Figure 5.14: Attacks by Abu Nidal and the PFLP while supported by Saddam Hussein (1990-1997)

Saddam faced three overlapping problems that he had to solve: 1) how to consolidate rule after the disastrous invasion of Kuwait; 2) how to prevent the military 185 from seizing control of the country via coup; and 3) how to prevent neighbors from capitalizing on the weakness of his regime. The provision of support to terrorist groups helped him solve all three problems and fortify his leadership once again.148

Concluding Remarks about Iraq While the support Saddam provided to terrorist groups during the run-up to the

2003 American invasion and the Iraq War will have to be examined in depth elsewhere, this case has demonstrated a pattern that has existed within the Iraqi state since independence. The leaders of Iraq have been subject to multiple coups and coup attempts and have structured the military forces so that they 1) were appeased by receiving a large share of the private goods generated by the oil revenue in the country and thus maintaining support for the regime; 2) balanced by creating parallel military forces and moving many of them from outside the control of the military itself; resulting in 3) a force that was large and eventually well-equipped, but unable to successfully defend against the external threat to the country’s leadership. The leader, in this case Saddam

Hussein, faced a level of High External and High Internal threat and, as the theory predicts, helped to reduce these threats to his regime by investing resources in terrorist groups that could weaken the enemy while keeping the threat to the leadership from the military at a minimum.

While there are additional predominant cases of state sponsorship (such as Hafez al ‘Assad’s Syria and Huq al ‘Zia’s Pakistan) that can, and should, be analyzed through

148 While the third threat period is reserved largely for future research, there are indications that Saddam’s increased support to Hamas and the Palestinian Islamic Jihad followed a similar pattern in the run-up to the Iraq war. 186 the lens provided by the Theory of Leadership Survival for a deeper understanding into the dynamics that result in state sponsorship, the final case considered in this study will be that of Mummar Qaddafi’s Libya.

MUMMAR QADDAFI’S LIBYA The final case of state sponsorship that will be examined here is that of Libya’s leader Mummar Qaddafi who, in his 40 years of power sponsored approximately 34 different terrorist groups all across Africa, Europe, South East Asia, and the Middle East.

Qaddafi’s Libya is a key case for the theory, not only because of the scale, intensity, and longevity of the sponsorship, but also because he was one of the few leaders to reverse this policy decision and essentially cease support for terrorism. Indeed, the longevity of

Qaddafi’s rule is phenomenal. Here is a leader that faced no fewer than 13 coup attempts during his rule, was attacked by the U.S. Air Force, suffered through numerous sanctions, defeat in international war, and ostracization by the Arab community. Yet he was one of the longest serving leaders in the contemporary world, retaining power for over 40 years.

Long considered an eccentric man with tendencies towards the flamboyant and bizarre, Qaddafi’s support for terrorist groups was often attributed to what was perceived in the international arena as his instability, irrationality, and reckless foreign policy agenda. Indeed, his foreign policy goals in general were often considered “a direct outgrowth of his personal ambitions and ideology.”149 Yet, perhaps a more useful explanation was given by Neurberger who stated, “the survival of Qadhdhafi’s regime

149 For a more nuanced perspective on the relationship between his personality and foreign policy see Monti-Belkaoui 1996, especially the section “Understanding Qaddafi” pp. 18-26. 187 was an important factor in shaping Libya’s foreign policy” (1982, 60) and when viewed through the lens of the Theory of Leadership Survival, his support for terrorism becomes more than a factor and rather, a rational response to the high level of external and internal threats faced by his regime.

There is no doubt that the immediate popularity of the revolution, the price of oil, and the investments in Libyan development left Qaddafi with a wide popular mandate and no small amount of ambition to remake the region according to a vision of Nasserist pan-Arabism and Islamic revivalism.150 In this light, St. John claims that “state-sponsored terrorism has often been the instrument of political leaders with ambitions in excess of their power base” and argues that this is clearly true in the case of Libya (1987, 48).

However, this statement, which is nearly an axiom in the study of state sponsorship, places the emphasis on the “ambition” of the leader, rather than problematizing the power base.

The Theory of Leadership Survival seeks to look more deeply into this shortcoming of Qaddafi’s power base and give some context to the insight that “cowing opposition to Colonel Qaddafi at home and abroad (wa)s one aim of Libya’s terrorist network” (Blundy & Lycett 1987, 180). This section aims to demonstrate not only that there is a correlation between the external mobilization of the military, the coup threat, and support for terrorist groups, but also that there is evidence that supports the claim that this sponsorship assisted Qaddafi in alleviating these threats.

150 Manifest in his Green book. 188 External Threat Environment in Qaddafi’s Libya There are three elements to the external threat environment for Libya. First, the anti-imperialist (especially anti-British) sentiment that was necessary for Qaddafi to consolidate his regime steadily increased tensions between Libya and the West, leading to a number of diplomatic and military confrontations, allegations of orchestrated coup attempts, and an attack Libya by the United States. Second, Qaddafi came to power in the context of a territorial conflict over a strip of land between Libya and Chad that had been raging since independence and which steadily expanded to include French and American intervention. Finally, relations between Libya and Egypt deteriorated leading to a number of confrontations, subversive actions on both sides, and eventually a four-day conventional war.

Whether or not these external threats were created by Qaddafi’s unbridled adventurism, there are a number of unanswered questions about his foreign policy choices that have been written off as mere “irrationality.” Indeed, it is unclear how an

“irrational” leader could have managed to hold on to power in an oil-rich state in a tremendously unstable part of the world for nearly 40 years. Rather, a more likely explanation is that Qaddafi found himself caught in a quagmire of needing to demonstrate his revolutionary credentials to the rest of the Revolutionary Command Council and reinforce his anti-imperialist and anti-Zionist image in order to sustain key domestic support for the new revolutionary regime. However, it wasn’t long before these foreign

“enemies” began to pose an actual threat to his regime.

189 Prior to the coup, Britain still enjoyed tremendous power and influence over the

Libyan government. Having taken control from the Italians following WWII, the British had granted Libya nominal independence, but remained very involved in every aspect of

Libyan security and foreign policy. King Idris had begun making moves towards anti- imperialism--a move that was very popular with the people—ensuring that any successor, or usurper, would continue this path (Ronen 2008).

When Qaddafi took power within the RCC, his anti-British stance helped assure the masses that this was an anti-imperialist coup, much like the Free Officers in Egypt.

Interestingly, the US initially supported the coup, and even tried to court Qaddafi, seeing him as a malleable and rabidly anti-communist ally (Blundy and Lycett 1987). It wasn’t long, however, before Qaddafi’s economic development projects led him to try and secure decisive shares in US oil companies--nationalizing them if refused—resulting in deteriorating relations with the Americans. As the relationship broke down, Qaddafi began to suspect that the CIA was working behind the scenes to remove him from power

(Ronen 2008; Blundy & Lycett 1987; St. John 2008).

A second front of external threat existed in Chad. Libyan interests in Chad were varied, vital, and bolstered by the surge in oil prices in the early 1970s. Qaddafi first invaded Chad in 1973, annexed the Aouzou Strip (which housed massive mineral, including uranium, and oil deposits)151 in 1975 and waged both direct and proxy wars there until Libyan affiliated troops were defeated in 1987. While Qaddafi clearly

151 However, Monti-Belkaoui & Riahi-Belkaoui argue Libya’s fundamental interest in Chad was historical and cultural rather than strategic (1996, 56). 190 benefited from the seizure of natural resources in both Libya and Chad, “Libya’s military leaders appeared determined to exploit the ongoing civil war in Chad” despite the fact that it was continuing to provoke the US and eventually led to direct US intervention which was further interpreted by Qaddafi as a “direct threat to Libya’s security,” and more specifically, to his regime (Ronen 2008, 23).

The conflict in Chad also affected the Libyan-Egyptian conflict, which erupted into a four-day long conventional war between Libya and Egypt in 1977. Much like the situation in Lebanon where both Syria and Israel worried that Lebanon could be used as a corridor for attack, Libya and Egypt became serious rivals who worried about the use of both Sudan and Chad in a similar way. This border dispute escalated to the point that it was a major military venture for Libya and there are indications that as early as 1982 the

CIA was involved in the conflict in Chad on the side of the government (Monti-Belkaoui

& Riahi-Belkaoui 1996, 59 & 60) and with Chad being a former French colony, France became increasingly involved as well.

When Qaddafi began to align with the Soviet Union in 1977, he dramatically increased military spending and equipment acquisition, but oddly did not increase the size of the regular military forces on par. For example, while the number of combat aircraft held by the Libyan air forces increased from 22 in 1972 to 555 in 1982, the number of personnel increased by only 2000 airmen. Indeed the ratio of air force personnel to combat aircraft plummeted over the course of the war with Chad.

191

Figure 5.15: The ratio of air force personnel to combat aircraft in Libya (1971-1990); Source: IISS 1971-1990

One explanation for this failure to increase air forces could be that while Army forces were necessary to fight in Chad, and aircraft were powerful symbols of power projection, piloted aircraft pose a direct threat to the seat of power in Tripoli.152

The war in Chad heated up just as the Libyan economy was crashing and the US,

France, and West Germany had moved in to support Chad. The military leadership began to staunchly oppose sending troops back into Chad, concerned that the conflict was straining Libya’s military capabilities and was becoming increasingly unpopular (El-

Khawas 1986, 111).153 As tensions with the US and Britain were reaching a breaking point--the US-Libyan rivalry having been exploited by both Reagan in the US and by

152 Interestingly, the years of the lowest personnel to combat aircraft ratio are precisely the years when Qaddafi’s sponsorship was the highest. 153 According to El-Khawas (1986) the military was opposed to Qaddafi’s foreign policy in general. 192 Qaddafi in order to bolster their domestic standing and increase their power—the conflict with Egypt showed no signs of abating and Libyan forces began to push back against deployment to Chad. This created a severe crisis for Qaddafi who began to believe that

“the powerful global alliance of the imperialist United States, Zionist Israel, and Arab- reactionary Egypt, Sudan, and Saudi Arabia was poised to take action against Libya”

(Ronen 2008, 23), a notion that the American “reprisal” attack did little to dispel.

Oddly, the inherited conflict with Chad over the Aouzou strip and Western support for the government in Chad rarely enter into scholarly analysis of Qaddafi’s decision to sponsor terrorist groups. On the surface, perhaps it appears that Qaddafi’s adventurism was nothing more than oil-funded megalomania, yet, “his polices (we)re deliberate and intentional, focusing on a) preserving his regime, b) expanding the influence of Libya, c) promoting Islam; and d) realizing Arab unity” (Monti-Belkaoui &

Riahi-Belkaoui 1996, 25). The competition over Chad, a vital interest to Qaddafi, played out in the international scene. This alone, however, was not enough to encourage a decision to sponsor. Rather, it was the international combined with the domestic level variables and issues facing the regime that pushed him in this direction.

Internal Threat Environment and Qaddafi The external threats Qaddafi faced were largely a consequence of his policies, but they were also the results of Qaddafi’s attempts to deflect internal threats. Qaddafi came to power following the One September Revolution in 1969, which was a bloodless coup orchestrated by the Libyan Free Unionist Officers to depose the Western-backed King

Idris. He quickly stepped into the spotlight as the face of both the new Revolutionary 193 Command Council (RCC)--consisting of all the leaders of the coup --and the Libyan state. Qaddafi’s first concerns were to consolidate power within the RCC, then the army at large, and then the country as a whole.

Again, it is worth mentioning that in new revolutionary regimes, there is often competition among the leadership to demonstrate their revolutionary credentials, which can lead to seemingly irrational foreign policies (Walt 1996). Although the nominal face of the revolution, in early 1969 Qaddafi still had to prove to the rest of the 12-member

RCC that he was not only the most revolutionary, but also had to find ways to solve mounting disagreements “over priorities, jobs, and other matters” within the RCC

(Cooley 1982, 84). By the late 1970s, Qaddafi had consolidated power, was able to leverage the economic situation, and demonstrated unequivocally that he was committed to the revolution. It wasn’t long thereafter that he dismantled the RCC completely.

The RCC was not the only source of immediate threat to Qaddafi’s rule, however.

By the early 1970s, Libya’s armed forces had tripled in size, expenditures on military equipment had continued to soar, and coup attempts became commonplace as Qaddafi’s policies continued to threaten the military establishment, at one point going so far as to try and dismantle the hierarchy completely (St. John 1987). Thus, despite purging the officer corps and raising salaries for the military across the board the first coup attempt against Qaddafi was mounted in December 1969, less than 5 months following the fall of

King Idris (Monti-Belkaoui & Riahi-Belkaoui 1996). By July 1970, less than a year after the revolution and other plot, known widely as the Hilton plot, was uncovered, and a third coup was attempted in July 1975 by Major Meheishi (Marshall & Marshall 2010). 194 By this time, Libya was bogged down in Chad, was facing deteriorating relations with Egypt, and Qaddafi had begun face opposition to his foreign policy from the within the RCC. These developments “coincided with Qaddafi’s (1977) shift toward closer ties with the Soviet Union, escalating expenditure on armaments, and increased support for revolutionary groups…”(El-Khawas 1986, 103). The RCC worried about a decline in world oil prices, pushed to reduce the amount of the national budget that was being spent on arms, the military’s commitment in Chad, and argued instead for new investments in social and economic programs (El-Khawas 1986).

This RCC push-back, and the 1975 coup attempt, were major factors in the dissolution of the RCC (St. Johns 2008) and the creation of the Revolutionary

Committees in 1977 which were designed as “shock troops for the Libyan revolution” that provided Qaddafi a way to purge counter revolutionaries and counter opposition to his rule (Monti-Belkaoui & Riahi-Belkaoui 1996, 23). These Revolutionary

Committees,154 though heavily armed in the service of protecting the regime, also bypassed the traditional military structure and answered only to Qaddafi himself. When

Qaddafi uncovered an assassination plot against him by the head of military intelligence and a key senior army officer in 1978, he further bolstered the Revolutionary Committees and ensured they bypass military chain of command occurred (Blundy & Lycett 1987). In an additional regime security measure, beginning in 1979, cabinets were regularly reshuffled in order to thwart attempts within the government itself to consolidate enough

154 They also formed the core of the “hit squads” that operated outside of Libya to assassinate dissidents, and were notorious for quelling not only dissent, but also discussion within the people’s councils and general committees. 195 power to contest Qaddafi’s rule. By the 1980s, coup attempts were common, with six occurring between May of 1985 and May of 1986 alone (St. John 2008)155 demonstrating without a doubt that the Libyan army was the most significant threat to Qaddafi’s regime.

Overall, Libya follows a familiar trajectory: a new leader takes power and works to ensure the loyalty of the army and the inner circle, there is an economic boom from increased oil revenue that is put into a military build-up which then is put to work in

Chad, but the army is kept fairly weak overall as they keep mounting coups d’état against the leader. Determined to remain in power, his military becoming less and less supportive of the revolution, France, West Germany, Britain, and the US undermining his interests in Chad and occasionally trying to overthrow his regime, Qaddafi turned to terrorist groups for survival.

Terrorist groups and Qaddafi’s leadership survival Qaddafi’s leadership survival became contingent upon successfully balancing three threats: first, countering the threat coup d’état from the military, second, neutralizing the counter-revolutionary forces within Libya and abroad, and finally thwarting attempts by France to ensure the defeat Libyan forces in Chad as well as those by the US and Britain to bring down his regime.

155 See also "Egyptian report of attempted coup in Libya," BBC Summary of World Broadcasts, March 8, 1979 available on LexisNexis accessed December 12, 2012; "Report of attempted coup in Libya" BBC Summary of World Broadcasts, August 20, 1980, available on LexisNexis accessed December 12, 2012; "Egyptian Report of Attempted Coup in Libya," BBC Summary of World Broadcasts, July 16, 1981, available on LexisNexis accessed December 12, 2012; "Sudanese Report of Attempted Coup in Libya," BBC Summary of World Broadcasts, January 11, 1982, available on LexisNexis accessed December 12, 2012; "Military coup attempt reported in Libya," BBC Summary of World Broadcasts, May 15, 1982 available on LexisNexis accessed December 12, 2012; “Attempted Coup'' Against Qadhafi" BBC Summary of World Broadcasts, February 28, 1983, available on LexisNexis accessed December 12, 2012; "Kaddafi Escapes a Coup" Newsweek, May 21, 1984, available on LexisNexis accessed December 12, 2012. 196 While the counter-revolutionary threat, was managed by the Revolutionary

Committees, both internally through exerting extensive control within the People’s

Committees and externally by use of hit squads drawn from the RC cadres, the other threats were countered by the use of external terrorist groups, primarily in Europe and the

Middle East.

There are indications that as early as 1969, Qaddafi was enlisting Palestinians operating on Libyan soil in the training of Chadian insurgents (Neuberger 1982,) and by

1972 was sponsoring nine active terrorist groups156 that were responsible for 333 attacks, several of which were Palestinian. Oddly, in July 1973, despite what appeared to be a major falling-out between the PLO and Qaddafi—to the point where Qaddafi had the

PLO training camps in Libya shut down—he continued to maintain “links with the various Palestinian groups including Yasser Arafat” (Arnold 1996, 60). While many Arab leaders had mercurial relationships with Palestinian groups, there are several factors that turn the spotlight back on this relationship.

First, Qaddafi was not only an ardent rejectionist who refused to entertain any thought of political compromise with Israel, but he also consistently and publicly shamed other Arab leaders for their inaction on the issue. Second, awash in oil money, Qaddafi built up an enormous stockpile of military equipment, yet did not increase the size of his military forces157 that would allow him to actually use the equipment. Both these moves

156 By “active” I mean groups that were perpetrating terrorist attacks at the time of support. The nine supported groups were the IRA, PFLP, MNLF, ETA, PLO, JRA, RAF, Black September, and the Tupamaros. 157 Several years later, he would attempt to dismantle the military completely in favor of a People’s Militia. 197 allowed him to maintain the appearance of a strong and stalwart military leader that was ready to wage war with Israel at the drop of a hat. Yet, when the 1973 Yom Kippur war started with a surprise Egyptian-Syrian invasion, just months following the PLO eviction from Libya, Qaddafi provided the other Arab states with equipment and fuel, but did not commit a single troop to the fight (Focus on Libya 1989).158

Using time as an additional variable to mark critical points of threat facing the regime, we can examine how closely this case adheres to the predictions of the theory.

Similar to the threat environments of Ayatollah Khomeini and Saddam Hussein, Qaddafi faced multiple threats his regime from the military within the first few years of his rule as well as in the mid-1970s. He also had numerous external threats to check, the most prominent of which were initially from Britain but soon grew to include the US and

France as the conflict in Chad internationalized and relations with Egypt deteriorated to the point of military conflict in 1977.

Given the exiting threats to the regime, we would expect that there would be a spike in sponsorship in the early 1970s when Qaddafi faced multiple coups, pressure from the US and Britain, and deployments of troops into Chad. We would expect another spike around 1975-1978 during the years when Qaddafi not only faced yet another coup attempt, but an assassination attempt. In addition, it was during these years that he began to receive tremendous military aid from the Soviet Union, and engaged in armed conflict

158 In all fairness, the conflicts between Qaddafi and Sadat led to Sadat blocking him from the “war room,” however, the outcome of the war led Qaddafi to only increase his rhetoric against the other Arab leaders, touting himself and the only one able to truly champion the Palestinian cause. 198 with Egypt. Roughly, Qaddafi’s support to active terrorist groups and their attack patters does match up with these time periods.

Figure 5.16: The relationship between threat of coup (plots and attempts), MIDs, and number of sponsored active terrorist groups in Qaddafi’s Libya.

Source: Coups (Marshall & Marshall 2010); MIDs (COW dataset); sponsored groups from the author.

What is most interesting is the downward trend in sponsorship from 1973-1974 when Qaddafi was focused primarily on domestic projects and implementation of his revolutionary agenda, oil prices were high, and much of the region, and the superpowers, were focused on the fall out from the 1973 war with Israel. In 1975, Qaddafi faced an

199 additional coup attempt from his military and his support to active terrorist groups also jumped back up to early 1970s levels.159

This coup attempt was what led him to completely reorganize the military, dissolve the RCC and instigate the Revolutionary Committees program, and shuffle the cabinet to prevent further threats to his rule. There was another jump in supported groups again in 1977 which coincided with these moves in addition to Qaddafi’s realignment with the USSR. This was also around the time when Libya began receiving significant military aid and equipment that strengthened the military tremendously—a military that was then deployed in a four-day campaign against Egypt--again raising the threat of the military to Qaddafi’s rule.

The dynamics in the 1980s were slightly different than the 1970s, but the trends in support and attacks remain roughly the same. Groups that attacked Qaddafi’s main threats in Chad and who targeted US troops and the British government remained those with key funding. There were indication as early as 1982 that Qaddafi was training the

Kanak Socialist National Liberation Front, the key political group seeking independence for the territory of New Caledonia held by the French in the Pacific (Mickolus, Sandler &

Murdock 1989) and the US State Department later disclosed that the Libyans were providing funding as well military training to the group “as part of Libyan opposition to

French policy in Chad” (St. John 1987). Furthermore, in 1989, the Libyan bomb planted in French UTA flight 772 that blew up over Niger on its way to Paris was deliberately

159 The IRA, PFLP, DFLP, MNLF, ANO, PLO, and Polisario Front 200 slated to be in retaliation for the French support of the Chadian government (Ronen

2008).

France, however, was not the only country intervening on the part of the government of Chad. The US was heavily involved by the early 1980s and Qaddafi was suffering heavy losses, some of which were responsible for the 1975 coup against his regime. Qaddafi stepped up attacks against groups targeting US military forces, including the New People’s Army in the Philippines and groups in Europe, culminating in a 1986 attack in a discotheque in Berlin targeting American and NATO soldiers for which

Reagan mounted a retaliatory attack.

While there is an ongoing debate regarding the utility of strikes vs. sanctions

(Collins 2004) in what finally persuaded Libya to give up support to terrorism, this theory lends weight to the idea that as the Soviet Union collapsed and Islamist groups took up arms against regimes across the region, Qaddafi no longer could leverage his support for terrorist groups as a way to protect himself against the convergence of external threats and the internal threat posed by the Libyan military. It is in this theoretical framework and Qaddafi’s policy reversal that we find the key to solving the puzzle of state sponsorship.

CONCLUSION All of these well-studied, prominent cases of state sponsorship have very similar domestic and international incentives and constraints that fall within the preview of the

Theory of Leadership Survival. All three regimes faced tremendous external threats, 201 some of which their legitimacy was hinged upon, as well as threats against their regimes from their standing militaries. Each saw an opportunity to help fortify their regimes by employing a terrorist group in their confrontations with external enemies, which allowed them to meet their challenges and maintain their legitimacy, all while keeping a significant portion of their military strength out of the hands of the military itself. All three cases conform roughly to the expectations of the Theory of Leadership Survival.

However, there were also cases in the dataset that did not conform to the theory.

The next chapter details two outlier cases. One case, of Tun Abdul Razak’s Malaysia, that was coded LELI, but where he did provide support for the Moro National Liberation

Front, and another of Fujimori’s Peru, which was coded HEHI, but where he did not sponsor terrorist groups. Exploration of these outlier cases will allow for refinement of the theory, open paths for future research agendas, and provide critical insights for policy formulation.

202 Chapter 6: Outlier Cases

The previous chapter discussed in depth the cases that supported the Theory of

Leadership Survival and helped to confirm the results of the quantitative analysis.

However, the data also indicated that there were cases that did not conform to the theory.

In several outlier cases, the leaders either faced all of the constraints and incentives of an

High External & High Internal (HEHI) threat environment, yet did not sponsor terrorism,160 or conversely they enjoyed the freedom from external and internal threat that was associated with the Low External & Low Internal (LELI) threat environment, yet chose to provide support for terrorism.161 While the cases in the previous chapter were used to illustrate how the leader used terrorist groups to fortify their rule by diverting some of the state’s military capabilities outside of the military chain of command, these cases will be used to explore direct challenges to the theory.

CASE SELECTION The outlier cases selected for this chapter fit one of the two extreme categories

(HEHI or LELI) but had outcomes in opposition to what the theory would predict. They were initially vetted by the same criteria as for the last chapter examining the confirming cases, yet those selection criteria had to be adjusted given the make-up of the categories and the cases within them. In the last chapter, and initially in this one, cases were selected

160 Although some, such as Said Barre’s Somalia, with further investigation, showed evidence of support to other non-state actors that perpetrated attacks. This indicates that the theory should be examined for applicability outside of the leader-terrorist relationship specifically. 161 Again, HELI leaders are also likely to sponsor terrorist groups, though the length and intensity of sponsorship is expected to be lower than in HEHI cases. This will be explored briefly in the section on LELI sponsorship, but more extensive investigation will need to be undertaken in a different study. 203 where the leaders in question were in power for a minimum of 10 years. This minimum was intended to ensure some uniformity over the all the examined cases allowing the bulk of the variation to stem from threat environment. However, there were no LELI sponsoring cases that met the minimum ten-year criteria, while 30 cases of non- sponsoring HEHI leaders did.

The selection criteria for the sponsoring LELI cases, of which there were originally ten, was then relaxed to include only those cases where the leader had been in power more than one year (over 12 months), which reduced the number of LELI sponsoring cases to six, only half of which sponsored more than 2 years. Of those three leaders which provided more than two years of support, only one showed an indication of having sponsored at least half the years they were in power, the other selection criteria for examining sponsoring HEHI cases in the previous chapter, helping rule out leaders who provide support for groups they are somehow not aware are terrorist. This leader, Tun

Abdul Razak bin Hussein of Malaysia, provided active military support and safe haven to the Moro National Liberation Front for five out of the seven years he was in power

(Salehyan et al. 2009; San Akca 2009). While the other LELI cases will also be examined briefly below, there are two additional reasons to select Abdul Razak. First, the level and duration of support Abdul Razak provided for the MNLF is generally undisputed.162

Second, the military in Malaysia, unlike in the cases in the prior chapter, are clearly subordinate to the civil government--having no history of coups, or even high-level

162 Although some have argued that it was the governor of Sabah, Tun Datu Mustapha that was the man responsible for the support. 204 insubordination, whatsoever, allowing for a strong case to be built in opposition to the

Theory of Leadership Survival. The case of Malaysia’s Razak bin Hussein, therefore, will be examined in depth below as a non-conforming case of a leader that faced an LELI threat environment, yet provided support for terrorism.

The case selection for HEHI leaders that did not sponsor terrorist groups faced similar, though opposite difficulties. As with the HEHI sponsoring cases in the previous chapter and the LELI sponsoring cases above, the non-sponsoring HEHI cases initially had only one selection criteria: that the leader remained in power more than 10 years.

However, there were 30 non-sponsoring leaders from 26 countries that fit the criteria, approximately 17% of all HEHI cases.

If we revisit the Theory of Leadership Survival, it predicts that support to terrorist groups are most likely to occur when leaders face an external threat that must be mobilized against, but this requisite build-up of the military coincides with an extreme distrust of the military, operationalized as coup threat to the regime, prompting leaders to place a portion of their military capabilities onto terrorist groups and outside of the control of the military. Therefore, a selection filter was developed to rank leaders by the number of external threat observation years in the data in relation to years in power, expecting that leaders facing more external threat would be incentivized to build up their militaries more and increase the pressure to sponsor.

There were seven non-sponsoring HEHI leaders who had at least 50% of their observation years containing a High External threat were: Eduard Shevardnadze

(Georgia), Kaysone Phomvihane (Laos), Heng Samrin (Cambodia), Hee Park (South 205 Korea), Said Barre (Somalia), Alberto Fujimori (Peru), and Felipe González Marquez

(Spain).163

In order to select from among these seven leaders, an additional selection criteria were employed: the leader must have directly faced a coup threat, setting aside those leaders who experienced an HI threat based on their predecessor’s threat, reducing the number of leaders to three: Heng Samrin (Cambodia), Said Barre (Somalia),164 and

Alberto Fujimori (Peru).165

Of the remaining two, Fujimori was chosen to examine in depth. Although Heng

Samrin was in power for 13 years, nine of which he faced external threat, Cambodia was not positioned, following years of Pol Pot’s rule, to be an active player on the regional or international scene. In contrast, Fujimori was engaged significantly with international momentary organizations, regional organizations, and the United States, indicating a clearly active foreign policy. Because state sponsorship of terrorism is assumed by this study to be a type of foreign policy under conditions when the leader faced explicit and acute external and internal threats to their rule, Peru was chosen as the most appropriate

(hard test) case of an non-sponsoring HEHI leader with which to confront the theory.

The exploration of the two outlier cases below should help shed some light on the conditions under which this theory fails to render the expected policy decisions. It will

163 Another nine leaders had between 25% and 49% of their observation years under HE threat, while 14 (a full 46.6%) of non-sponsoring HE leaders had less than 25% of their observation years include an HE threat. 164 Said Barre was disqualified from examination based on the fact that while he did not support “terrorist” groups as defined by this study, he did actively support insurgent groups across Ethiopia during the Ogaden war and the purpose here is to locate a “hard case” to challenge the theory. 165 In Chapter 3, leaders were determined to face a HI threat if they themselves or the leader before them faced coup attempts or were overthrown by a successful coup. 206 also allow a closer look at alternative ways leaders react to HEHI pressures in order to retain power, which my not only lead to for further refinement of the theory and future research agendas, but to solid policy prescriptions for how to end, and prevent, state support for terrorism.

LELI CASES OF SPONSORSHIP The Theory of Leadership Survival predicts that few, if any, leaders facing an

LELI threat environment will provide support for terrorist groups. The purpose of this section is to explore in depth the case of Tun Abdul Razak bin Hussein of Malaysia who provided active military support and safe haven to the Moro National Liberation Front

(Salehyan et al. 2009; San Akca 2009) despite the low threat environment that both the absence of lack of MIDs (LE) and coup threat (LI) the data indicate.

MALAYSIA’S TUN ABDUL RAZAK BIN HUSSEIN Similar to Saddam Hussein in Iraq, when Tun Abdul Razak came to power in

September of 1970 he was already the de facto Head of Government (Shaw 1976) and had been responsible for restoring order via martial law from some of the most deadly and chaotic episodes of ethnic violence in recent Malaysian history. Having served as

Director of Operations over the course of several years, Abdul Razak was the clear choice to take the reigns when the prior leader stepped down to take the position of

Secretary General of the Islamic Secretariat in Saudi Arabia (Shaw 1976). Abdul Razak took power at a critical time in Malaysian history and political development. It had fully consolidated as a country, after having separated from Singapore, was moving away from

207 British protection, and beginning to look outward to define its role in the world as an independent state. Demographics and the ethnic based power structure in Malaysia incentivized the government to step out onto the global stage as a Muslim nation, pushing for membership with the Organization of Islamic Countries (OIC), and reorienting

Malaysia sharply away from what had been British protection and toward regional alliances and identification with the non-aligned movement (Ministry of Foreign Affairs,

Malaysia).

There were several critical domestic issues that demanded Abdul Razak’s attention: first, the long-standing territorial dispute over the Malaysian state of Sabah in

Northern Borneo between Malaysia, Indonesia, and the Philippines. Second, the ethnic tensions and violence within Malaysia erupting only intermittently, but with such force, that it threatened to destabilize the country. Finally, the communist insurgency raging within Malaysia itself that posed a threat to Abdul Razak’s hold on power.

The Theory of Leadership Survival predicts that leaders with low external threat will have little need to mobilize their militaries for battle and when this is combined with a low threat from the military itself, there will be little, if any need to pursue relationships with terrorist groups. The case of Tun Abdul Razak’s providing safe haven, arms, and explicit military support to the Moro National Liberation Front (MNLF) a Muslim separatist terrorist group that was, at the time, fighting the government of the Philippines for the independence of Sulu and Mindanao in the Philippines as well as Sabah in

Malaysia, then, presents a direct challenge to the theory, necessitating a deeper look into

208 the dynamics of the civil-military relations, the external threat environment, and the relationship between Tun Abdul Razak and the MNLF.

External Threat Environment Tun Abdul Razak inherited a state that was weaker militarily than his immediate rivals and neighbors, Singapore, Indonesia, and the Philippines, in part because

Malaysian security had been guaranteed by the British army since independence. Now with a general retreat from the region by the British and the Americans following suit in

Vietnam, Abdul Razak decided to push for a policy of neutrality in his immediate neighborhood, wagering that non-alignment, neutrality, and increasing alliances with regional powers, such as China, would keep Malaysia relatively safe from regional conflicts and the proxy battles of the Cold War.

Although throughout Abdul Razak’s rule, the region was wracked with instability, communist insurgency, foreign intervention, and cross-border military disputes, Malaysia experienced no active MIDs and mobilized no forces for external defense, thus characterizing Abdul Razak as a leader facing a low level of external threat for the purposes of this study.166 However, despite the lack of military deployment, Malaysia was undeniably engaged in a deep rivalry over the region of Sabah that involved both

Indonesia and the Philippines and in fact, over the course of Abdul Razak’s tenure in

166 However, the leader immediately prior to Abdul Razak, Tunku Abdul Rahman Putra Alhaj, experienced an MID in 1968 when Abdul Razak was acting Head of Government. 209 power, Malaysia invested a greater percentage of its GDP in its military than any of its rivals.167

Figure 6.1: Military Expenditure as % of GDP when Abdul Razak was in power

Indonesia, in the early 1960s, well before Abdul Razak took power, instituted a policy known as “Confrontation” against Malaysia, which intentionally stopped short of military conflict, but did include troop movements and other intimidating postures along their border with Sabah. At this time, Malaysia was still generally under the protection of the British, with many British, Australian, and New Zealander troops stationed there

(Enloe 1978). Indeed, despite the Indonesian intimidation and Malaysian resources that were reappropriated to military spending, because there was no “full-scale attack,

167 Interestingly, this expenditure level is consistent with what the theory predicts in terms of military expenditure and terrorist group support, even though the measurements of the external threat are not. 210 Malaysia was saved from making any drastic, as opposed to sizeable, increases in her armed forces” (Milne & Mauzy 1978, 302). Thus, while Malaysia had to keep an eye on

Indonesia, and while the reversal of this policy was greeted with “relief” by the

Malaysian government (Milne & Mauzy 1978, 303) the policy of confrontation was not treated as an acute threat.

Abdul Razak’s policy of regional neutrality in order to keep Malaysia out of conflict struggled to get off the ground in large part because Filipino leader Ferdinand

Marcos refused to concede the claim to Sabah (Shaw 1976). In fact, although Marcos worked to reestablish friendly relations with Malaysia his first years in power, he simultaneously trained a covert unit of approximately 30 Filipino Muslims to infiltrate and take Malaysian-held Sabah. When the group became insubordinate—stories vary as to whether it was over failure to receive their pay or refusal to invade Sabah—they were all killed (Nobel 1976), sparking the formation of widespread rebellion against the

Filipino government and of the MNLF (fas.org). Because of its close proximity to the regions of the Philippines for which the Moros were fighting to gain autonomy, Sulu and

Mindanao, the region of Sabah was the focal point for Malaysia’s ongoing conflict with both Indonesia and the Philippines and, perhaps not coincidentally, was the base location of the MNLF.

If the external threat to Sabah in particular and Malaysia in general was not acute, but rather a chronic low-level irritant, was it enough to influence Abdul Razak to condone the long-term sponsorship of the MNLF or did the context of the domestic

211 environment play a part? In either case, what are the implications for the theory of leadership Survival?

Internal Threat Environment Abdul Razak came to power during a time of general unrest and ethnic strife bordering on civil war, yet the military itself was never a threat to the regime. Since

WWII, Malaysian civil-military relations have been characterized as one of

“uninterrupted civilian control” (Alagappa 2001, 435). The threat of coup d’état was not something that the regime seemed to take into account when making policy decisions, therefor the internal threat from the military is correctly understood as low, as the data indicate.

Although the Theory of Leadership Survival does not specifically address the effects of civil war and ethnic conflict on a leader’s decision to sponsor terrorism, there is an abundance literature that takes this into account and understands it as an additional source of internal threat to the regime, which the results of the quantitative analysis support.168

Malaysia is an ethnically organized society with ethnic Malay political dominance protected by the right-wing United Malays National Organization (UMNO), the political party to which Abdul Razak belonged. The Malay hold key positions in the military and police forces, and ethnic Malay are overwhelmingly represented in the military (Enloe

1978), as well as in government, giving them disparate control over how resources are

168 However, in this case, the Political Instability Task Force (PITF) data from which the internal war variable is sourced codes the MNLF as an ethnic war that is taking place in the Philippines. 212 distributed. Not only did Malays hold key positions in the military, but also predominated the rank and file and the “political leadership came to view the military and the police force as a crucial though not sufficient basis on which to maintain Malay political dominance” (Alagappa 2001, 443). Thus, unlike the cases of HEHI sponsorship explored in the previous chapter, Abdul Razak faced no threat from the military rather he utilized the institution to maintain the status quo.

The key points of instability that challenged the Abdul Razak regime were the same that had threatened his predecessor and had ushered Abdul Razak into power in the first place. Ethnic violence, especially around elections, were explosive and there are indications that the Sabah conflict was exploited by Abdul Razak to provide a rallying point for an electorate otherwise consumed with racial and ethnic hatred, and in addition, there was a resurgence of the communist insurgency in Malaysia in 1975 (Shaw 1976) that was at least in principal supported by China. However, the threat faced by Abdul

Razak was neither explicitly from the military nor from an acute Militarized International

Dispute.

Malaysian Support for the MNLF The support provided for the Moro National Liberation Front in Sabah by the

Malaysian regime was long lasting, extensive, and blatant. The regime provided arms to the MNLF between 1973 and 1976 (San Akca 2009) and “explicit military support” began in 1972 (Salahyan et al. 2009). There are also indications that Tun Datu Mustapha

(the chief minister of Sabah) used his position to provide financial support to the MNLF

213 (Miller & Smarick 2012) and that the Malaysian government intentionally did not interfere with his provision of this support (Noble 1976).

Although the Global Terrorism Database only records seven attacks by the MNLF during Abdul Razak’s tenure in power--five instances of hostage taking and two of armed assault--they have been overall responsible for at least 187 attacks all but 17 of which were against non-military targets.

The MNLF sought support from groups like the Organization of Islamic

Countries and Muslim leaders such as Sabah’s Chief Minister Mustapha and Libya’s

Qaddafi are known to have been some of their biggest supporters, yet the “Malaysian involvement in supporting the Moros was not so much based on their commitment to religious duty as to pressure the Philippine government to drop its claim on Sabah”

(Samad & Bakar 199, 560). This territorial dispute was further escalated by the fact that the Filipino army redeployed 70%-80% of its forces to the south in order to counter the insurgent threat posed by the MNLF.

WHAT EXPLAINS MALAYSIAN SUPPORT FOR THE MNLF? If Abdul Razak faced neither an internal threat of military coup nor the an acute

Militarized International Dispute as the Theory of Leadership Survival expects in order to explain support for terrorist groups, what does? More importantly, what are the implications for the theory?

First, with extreme caution regarding stretching the theory to include cases that do not fit the carefully defined criteria, the exploration of this case has made it clear that the

Malaysian government did not exactly inhabit an environment of “low” external threat 214 and the military expenditure as percent of GDP reflects this reality. While the long- standing dispute over Sabah between Malaysia and its neighbors never escalated to conventional warfare, or even mobilization for such, policies such as the Indonesian

“Confrontations” and Marcos’ training of subversive fighters to infiltrate Sabah clearly indicate a level of heightened external threat.

Without changing the official designation of LELI, it is still possible to accept that

Abdul Razak faced a certain level of an HELI threat. When examined hypothetically as an HELI case, the support provided to the MNLF falls well within the bounds of what is expected for HELI leaders. First, the shadow of the failure of previous regimes are acute influences in policy decisions; second, the military is not expected to act as a threat, but rather as a noncompliant force that either will not deploy to meet the external threat because of policy divergence with the leader or cannot because of real constraints on forces (nuclear arms, internally focused military dealing with an insurgency, democratic pressures to pursue nation building at home, etc.). In these cases, leaders are expected to utilize terrorist groups only as long as they fully work to meet the policy needs of the leader. These leaders are not caught in the same sort of “strategic bind” (Byman 2005) that HEHI leaders face.

Again, while not reclassifying this case, the support for the MNLF seems to be best explained by a combination of preemption against a specified internal threat—trying to rally various ethnic groups against an external foe in preparation for elections—and in response to mass troop redeployment in the south by the Filipino army.

215 WHAT ABOUT THE OTHER LELI CASES? The other nine LELI cases of sponsorship will be examined briefly here. First, there were three cases in France, Charles de Gaulle, Georges Pompidou, and Alain Poher.

The case of de Gaulle is difficult to examine in depth here because there were only two out of his nearly eleven years of his rule represented in the data set (1968 and 1969). Both years were coded as supporting terrorism, however, there are indications that he provided what was probably closer to “passive” support to the Basque separatists, ETA. Although the details of his support are not entirely clear, it is most likely that he merely did not act to eject the ETA from the Basque lands in France rather than providing true active support, which is the focus of this study.

Two other French leaders show indications of supporting ETA, first was Alain

Poher, who served only two months as Acting President in 1974 and thus cannot truly be considered a sponsor of any kind and second, Georges Pompidou, who preceded him.

Pompidou allegedly provided sanctuary for ETA in French Basque territory (“How to be a Basque on both sides of the border” 1979), in part because he understood ETA to be a threat specifically to Franco’s rule. However, he also allowed many other groups, including the Japanese Red Army to open offices in Paris (Sterling 1981). Thus,

Pompidou is another case that should be examined in depth as an LELI sponsor of terrorism, although the total amount of support was only approximately 40% of the time that he was in office.

The case of Latifur Rahman in Bangladesh is similar to that of Poher in France, he served as Prime Minister from July to October 2001, and it is therefore quite unlikely that

216 he was responsible for the policy of sponsorship. However, both his predecessor, Hasina

Wazed, and successor Khaleda Zia faced a LEHI threat environment and were likely responsible for the safe haven and training camps provided for multiple Indian terrorist groups: United Liberation Front of Assam (ULFA), the National Liberation Front of

Tripura (training and expertise) the National Democratic Front of Bodoland (NDFB) (San

Akca 2009; UCDP external support dataset). The LEHI cases of state sponsorship have been under examined in this study despite the prominence and importance of the cases and will be explored in future work, as will be outlined in the following, concluding, chapter.

Khaleda Zia likewise served as Prime Minster of Bangladesh from 2001-2006 and the only year available for analysis in this data (2001) was also coded as LELI. Because of the truncated data, it is difficult to clearly define her support of various Indian groups as having occurred in an LELI environment, especially since her previous administration sponsored during an LEHI one. Furthermore, the building of the security fence across the

Bangladesh-India border is indicative of existing tensions between the two nations.

The case of Sabah As-Sabah of Kuwait is also complicated in the sense that the group he supported was the PLO. Unlike some of the factions of the PLO, the overall umbrella organization was not considered to be terrorist, but rather to be fighting Israeli neocolonialism and much of the legitimacy of the conservative monarchies in the region at the time was dependent on their support for the Palestinian cause. While not discounting the actual damage done by the PLO, the regional context should be taken into account. In other words, during the 1970s, there were no regional Middle Eastern powers 217 that did not support the PLO and therefore, this case does not necessarily constitute a strong challenge to the theory.

The final remaining LELI sponsoring cases are that of Veselin Djuranovic who, again, was only in power a very short time May of 1984 to May of 1985 and Radovan

Vlajkovic who served a similar term from May 1985 to May 1986. It is unclear which of these men were actually responsible for the support as the only real information available is the allegation that, by 1985, Abu Nidal had “a considerable base” in Belgrade (Seale

1992, 37) and that “Yugoslavia let Abu Abbas, PLF leader accused of masterminding the

Italian cruise ship hijacking, flee the country, even though the U.S. had issued an arrest warrant for him” (Sabasteanski 1990, 309). Furthermore, it is difficult to attribute any of these developments specifically to these two men, given that ties to terrorists, such as

Carlos the Jackal and other groups, were actually formed under the Tito administration

(Yallop 1993, 443; Sterling 1981, 147).

The relationship between the decision to begin sponsorship and to continue sponsorship is not yet well understood. However, there are indications that one of the biggest predictors of a leader’s propensity to support terrorist groups is whether the leader before them did. Given this tendency to continue certain institutionalized policies,

LELI sponsorship should not be taken out of the broader context of the sponsoring history of the state and this type of sponsorship should have the most attention paid to it when it is not preceded by a higher level threat environment and/or previous sponsorship.

In summary, the bulk of the LELI sponsoring leaders were only in power for a very short time and had predecessors who faced a different threat environment than LELI 218 and instituted sponsorship as a policy. Those that did not fit this model, such as Sabah

As-Sabah in Kuwait, were in a situation where there was not only a regional expectation to support the group, but their internal legitimacy hinged on it. In other words, while none of these cases fit the model presented in the Theory of Leadership Survival, they also do not represent major outliers that cannot be explained.

HEHI NON-SPONSORING STATES Unlike the above LELI sponsors who did not face an external threat or internal threat, yet sponsored terrorist groups, this section explores the situation of HEHI facing leaders who chose a policy option other than state sponsorship of terrorist groups to fortify their regimes in power. The intention of this section is to explain how HEHI threat facing leaders were able to manage/dissipate the high internal/external threat and protect their positions of power without the utilization of terrorist groups.

ALBERTO FUJIMORI’S PERU A second outlier case examined in this chapter is that of Peru under Alberto

Fujimori who faced an acute HEHI threat, yet made policy decisions contrary to the expectations of the Theory of Leadership Survival, i.e. He did not extend support to external terrorist groups. Fujimori, elected to power in 1990 as a political outsider on a platform of anti-corruption and anti-terrorism, had a powerful ally in Vladimiro

Montesinos.

Fujimori found himself immediately warding off imminent domestic collapse. He inherited a country facing massive debt and “an overall inflation rate, accumulated over

219 the 5-year course of the preceding Alan Garcia government (1985-1990), of 2 million percent” (Palmer 1997, 115) not to mention a Maoist terrorist-insurgency by the Shining

Path and the Tupac Amaru Revolutionary Movement (MART), that had killed and wounded tens of thousands and all but destroyed the state infrastructure. In addition, an age-old boundary dispute with Ecuador flared up repeatedly and culminated in a conventional war in 1995. In these circumstances, the external and internal threats facing

Fujimori were beyond acute.

Although he faced a high internal and high external threat environment for a significant portion of his rule, Fujimori did not provide support for external terrorist groups. The Theory of Leadership Survival assumes that mobilization for external threat bolsters the resources available to the military, which then increases the threat of coup d’état that the military already poses to the ruler. In the case of Fujimori, he was a political outsider with no party base and, while popular among the people, many of his policies were highly divisive among the elites. The case of Fujimori’s Peru, then, is complicated for the Theory of Leadership Survival. Fujimori faced a clear HEHI threat, yet the internal aspect of the threat was most acute not from the military, but rather from a plethora of domestic actors that the theory does not explicitly address.169

Again, while being wary of stretching the theory to fit all existing cases, it is likely that the HEHI threat faced by Fujimori did not result in support to an external terrorist group because that action could not help him balance the threats he faced from

Ecuador, his military, the Shining Path and MART, and from the other branches of

169 Specifically insurgent groups, opposition in the legislature, and the judiciary. 220 government that threatened to bring down his regime. While external terrorist groups could not assist Fujimori in securing his regime against the multiplicity of threats,

Vladimir Montesinos, head of Peru’s National Intelligence Service (SIN), bought off judges, congressmen, the media, and high-ranking members of the military in order to shore up Fujimori’s power (McMillian & Zoido 2004). In addition, Ahram’s (2011) work on proxy militias indicates that leaders facing acute internal threats use domestic proxy militias and death squads to shore up the leadership – a similar argument to that made by the Theory of Leadership Survival regarding the use of external groups. Thus, while

Fujimori did indeed face an HEHI threat, Montesinos’ stealthy corruption rings and the use of internal death squads helped to mitigate the threats to Fujimori’s rule in a way that reliance on external groups could not.

External threat The territorial dispute between Ecuador and Peru was one of the longest running in the Western Hemisphere. The Rio Protocol of 1942 was supposed to have settled it once and for all, yet it flared once again into a full-scale conventional war under

Fujimori. The conflict originates in a long section of Peru-Ecuador border along the mountain range Cordillera del Condor that was never demarcated. The boundary settlement centered on the notion that there was a single watershed between Ecuador and

Peru that could help define the border in the remote area. However, an aerial survey by the US Army Air Force, brought in under the Rio Protocols to help enforce the settlement, discovered that there were, in fact, two watersheds. Given this discrepancy, the government of Ecuador began to use this “error” to argue against Peruvian claims to 221 the area and by 1960 had declared the Rio Protocol essentially null and void (St John

1996, 81).170

Tensions remained high, occasionally erupting into cross-border skirmishes and in

1981 flared up into a conflict that was just shy of a conventional war. Further cross- border clashes in 1991 led to a full-scale conventional war in 1995 taxing the cash- strapped and military-downsized Fujimori government even further (Cooper 2003).

Despite having dominated in earlier wars, the Peruvian military had suffered significant cuts under the Garcia government and had been redeployed in the early 1990s to the highlands to fight an insurgent war against the Shining Path. By the time an Ecuadorian helicopter bombed a Peruvian army post inside the Peruvian border sparking the 1995 war (“The 1995 Peruvian-Ecuadorian Border Conflict” 1995) the low level of preparedness, including air defense, radar, and available aircraft, put Peru at a significant disadvantage (Cooper 2003).

This lack of military preparedness, including weaponry, defense systems, and troops would be the perfect situation in which to expect a leader to employ terrorist groups in attacking the enemy, yet in the case of Fujimori’s Peru, this did not occur—and

Peru lost the war. Why did this external threat, and clear military inferiority, not lead to the use of external terrorist groups against Ecuador? According to the Theory of

Leadership Survival, the simultaneous internal threat environment is also expected to have an influence on leader sponsorship behavior.

170 The interest at stake being Ecuadorian access to the Maranon River, something that is not allowed under the Rio Protocol and explicitly rejected by Peru. 222 Internal threat Fujimori was democratically elected in 1990 on a platform of fighting corruption, stabilizing the economy, and crushing the Shining Path and MART insurgencies. Despite progress over the first couple of years, Fujimori began to run into significant opposition in congress, especially regarding some of the radical measures that he thought necessary to restructure the government and reintegrate into the global economy as well as to fight the Shining Path.

While the Peruvian military had a long history of not only praetorian rule, but also coup plots and attempts, it was not long before Fujimori had been essentially co-opted by a faction affiliated with General Nicolas De Bari Hermoza Rios171 who headed the

Counter-Subversive Internal Front (Huby 1994). Not much longer after that, Montesinos was able to buy off many of the remaining factions (McMillian & Ziodo 2004). In 1992, with the military’s help, Fujimori seized control of the government, dismissed congress and suspended the judiciary, bringing an end to a democracy that was only twelve years old, having been a military regime prior to that.

This coup, perpetrated by Fujimori and a faction of the military was known as an

“autoglope,” (self-coup) and was followed almost immediately by an attempt by other military factions to counter it. The officers indicted in this second coup attempt were subject to harsh treatment and even torture by Montesinos and his men, yet some discontent in the military continued to exist (“Peru: Military Unease Growing” 1992).

171 Who was also the chief of the general staff and commanded the army in general. 223 Internally, then, the threat environment was accurately coded as high. The threat from the military played out in the context of military coups both in support of Fujimori and then shortly after, to attempt and restore a democratic system. Given this convergence of High External & High Internal threat, the Theory of Leadership Survival predicts support for terrorism, yet there is no indication that Fujimori provided support to external terrorist groups. What then explains this outlier?

Keeping with the logic of omnibalancing on which this theory rests, a closer examination into the true sources of threat to Fujimori’s rule is warranted and there are indications that the threat posed by the Shining Path was a much bigger threat to

Fujimori’s rule than the Peruvian military. First, the military had been significantly weakened by Garcia’s shift of resources to the police, by serious cuts that had been part of dealing with the economic crisis, and by Hermoza’s restructuring of the higher echelon. Second, it is also clear that Montesinos had been paying off key figures in the

Peruvian military in order to keep their loyalty to Fujimori’s regime.172 Finally, the tremendous damage that the insurgency had been causing to the economy and the infrastructure of the state pushed the domestic insurgency, especially that waged by

Shining Path, to the top of the threat agenda.

Shining Path and MART: Insurgency in the Highlands Similar to the other outlier case of Abdul Razak examined here, the major threat to Fujimori’s regime came from insurgent groups throughout the country. By the time

172 Although this type of appeasement is present in the HEHI cases where the leaders did sponsor terrorist groups as well. 224 that Fujimori took power, Peru had already endured nearly 10 years of attacks from

Maoist groups like the Shining Path. Between 1980 and 1993 Peru experienced 5,880 attacks claiming almost twenty thousand lives (Huby 1994). Peru’s economic crisis was intricately woven into the insurgency waged by the Shining Path and the MART. While

MART was more tied into the “international terrorist” network, given Cuban and Libyan ties, it was the Shining Path that did the most damage to the Peruvian state, and posed the most threat to Fujimori, and thus will be the focus of this section.

Headed by philosophy Professor Abimael Guzman, the Shining Path adopted an ideology that was a mixture of a Maoist interpretation of socialism (and insurgency) and

Tauting, an Aztec-Christian mythology about life, sacrifice, redistribution, and the apocalypse (Strong 1992). With emphasis an on land redistribution, destruction of the centralized state, and religious explanations for their insurgency, the Shining Path became a key force among the peasants of the highlands, especially as the state became increasingly brutal in its attempts to militarily suppress the insurgency.

The rate of attacks by the Shining Path (and MART) during Fujimori’s rule were very high until Fujimori’s self-coup and Guzman’s capture in 1992, after which, they dropped off tremendously.

225

Figure 6.2: Terrorist attacks by the Shining Path and MART in Fujimori’s Peru (1990- 1999).

The counter-insurgency was fought in a number of stages, beginning well before

Fujimori’s rule. Initially, it was handled exclusively by the military, which put the affected provinces under emergency rule--effectively putting the administration of the providence under the purview of the military—and approached the fight as one of

“internal war” where the Shining Path was treated as an opposing army.

By the 1980s, under Garcia, this approach had resulted in tremendous civilian casualties, significant human rights violations, and very little headway against the insurgency. Additionally, given the brutality of the military against the various highland populations, the Shining Path was gaining support among people who thought they could provide a measure of protection against the government forces. A second approach was

226 attempted in the mid-1980s, one of development of the key areas of the fighting, following the rationale that the insurgency was popular primarily because it offered a vision of development and a way out of poverty for the people. The government before

1985 could not compete with this. This “developmentalist” approach likewise met with little tangible results.

By 1987, under fire for charges of human rights abuses, and suspicious of military power, Garcia had begun to shift resources to the police forces at the expense of the military (Strong 1992). When Fujimori took the reigns following his self-coup, he put the counter-insurgency largely under the jurisdiction of the security and intelligence services.

Citizen groups called rondas became common, were organized by a military commander to fight the Shining Path, and were far from voluntary, since refusal to cooperate would mark one as a sympathizer, with dire consequences (Mauceri 1995).

It is here, also that Fujimori becomes associated with the use of death squads, such as Grupo Colina (Barbier 2011), which both specifically targeted the insurgents as well as intimidated villagers to prevent them giving support to the insurgents. Fujimori has disputed how personally responsible he was for the actions of the death in international courts.173 It is undeniable, however, that it served his political purposes, provided the political legitimacy with some of the people, and with his party, to keep him in power, and simultaneously, increased the power Vladimiro Montesinos over the military, the government, the media, and over Fujimori himself.

173 Indeed, it’s likely that Montesinos orchestrated this program with the help of General Hermoza. 227 Why didn’t these threats and dynamics drive support for terrorism as it did in the regimes of Ayatollah Khomeini, Saddam Hussein, and Mummar Qaddafi? It appears

Fujimori’s case was different for three reasons: first, external terrorist groups could not mitigate the particular threats he faced. Second, even if they could have been, there were no available terrorist groups for him to harness. Finally, what internal threats outside of the insurgency he did face, were more than adequately handled by the corruption schemes of Montesinos. The co-option and buy-offs that he orchestrated were immensely intricate, well documented, and overlapping. When the buy-offs were finally exposed, Montesinos fell and took Fujimori regime with him. The following sections will examine these differences at length.

Why didn’t Fujimori support terrorism? Given what we know of his internal and external threat environment, the Theory of Leadership Survival expects that Fujimori would support terrorist groups in order to counter Ecuador and minimize the threat of coup by a mobilized Peruvian Army.

However, the theory would also expect that the groups Fujimori would support would somehow aid him in countering these threats and thus place to look for those groups would be in Ecuador itself.

Ecuador, according to the START database, is home to six terrorist groups: the

Armed Revolutionary Left, Ecuadorian Rebel Force, Group of Popular Combatants

(GPC), People’s Revolutionary Militias, Revolutionary Armed Corps (CAR), and White

Legion (Miller & Smarick 2012). The only group, however, that was active during the time that Fujimori was in power, and thus even available to support, was the Group of 228 Popular Combatants (GPC). The GPC is thought to have formed in 1994 as the military wing of the Ecuadorian Marxist-Leninist Communist Party (PCMLE) with suspected ties to the ELN and FARC.

While common ideology is not a prerequisite for state support to a terrorist group, and often leaders will support groups with divergent ideologies in order to achieve security, it would be a stretch to expect Fujimori to support the GPC for two reasons.

First, this was not just a case of divergent ideologies, but rather completely opposing one.

As Fujimori moved to restructure the Peruvian economy, he positioned himself squarely in the camp of the very policies that the GPC was fighting in Ecuador. As the Fujimori government moved further right, the likelihood that they would be open to collaboration with the GPC lessened.

This case does help illustrate one key point that perhaps the theory has not yet made explicit, that is, in order for leaders to enlist the assistance of terrorist groups, those groups must exist in the first place. In addition, this case shows that there are multiple ways in which leaders can fortify their regimes and the state sponsorship of terrorism is merely one. The following section examines how Montesinos was able to fortify the regime of Alberto Fujimori.

Montesinos Vladimiro Montesinos, a military officer who was trained at the U.S. Army

School of the Americas, convicted of selling secrets to the CIA in the 1970s and subsequently thrown out of the army and jailed, renewed his career as a civilian lawyer defending drug dealers. He first came into contact with Fujimori during the 1989 229 Presidential campaign in Peru when he defended Fujimori against allegations of fraud

(“Montesinos: The end of the road” 2001).

Fujimori came to power with the aid of Montesinos and was more than dependent upon him to remain in power. Following the 1992 autoglope, Montesinos helped Fujimori consolidate power, silenced the opposition through bribery and blackmail, and expanded the counter-insurgency operations in the highlands against the Shining Path. This included the widespread use of death squads and quickly resulted in the capture of their leader, Guzman.

Despite the precise, and nearly complete, nature of the blackmail, an independent television network aired a videotape in 2000 of opposition congressman Alberto Kouri accepting US$15,000 from Montesinos and agreeing to switch sides to support Fujimori.

Before long, other videotapes aired and it was estimated that more than 1,600 people in congress, the judiciary, the military, and the media were being paid off by Montesinos to support the president (McMillian & Zoido 2004). Interestingly, Montesinos kept careful records of each transaction, video or audio taping most of them and keeping written records as well. When once the first videotaped aired, Fujimori reportedly began to panic and threaten Montesinos, who responded by saying, “If necessary, I can set the prairie on fire” (McMillian & Zoido 2004, 22). Fujimori, left with no way to preserve his power, fled to Japan.

CONCLUSION Both Malaysia and Peru provide instances of outliers that do not conform to the theory, yet both cases also work to shed light on the choices that leaders make as they 230 work to balance the internal versus external threats to their rule. Fujimori faced an external threat from Ecuador that erupted into fighting more than once and faced a coup attempt, yet his biggest threat came from the Shining Path and the threat of economic collapse. Montesinos bolstered this regime, oversaw the death squad and other counter- insurgent operations in the highlands that brought the Shining Path to its knees, furthering

Fujimori’s legitimacy and reputation as the man that could put Peru on the path to greatness. In short, there were no external terrorist groups that could assist Fujimori with maintaining power and decreasing the threats to his rule.

Abdul Razak in Malaysia on the other hand, found the MNLF very helpful in maintaining stability and security in his state. While the threat environment was technically LELI, it was clear that the Philippines threatened to take control of Sabah and this threat actually provided a rallying point for the Malaysian people, otherwise mired in ethnic conflict.

In sum, while there are cases that do not conform to the theory, they do so in ways that continue to shed light on the motivations of leaders to sponsor groups that target civilians. In nearly every case, the leaders made calculations that allowed them to meet their objectives while foremost preserving their positions of power. In fact, it was this preservation of power that helped define what their threats and objectives even were. The final chapter will examine the implications of this theory and further lay out the conclusions of the study as a whole.

231 Chapter 7: Implications and Future Research The Theory of Leadership Survival developed and tested in this study sheds new light on the political logic of state sponsorship. Specifically, it moved away from a strictly outside-in or inside-out explanation of state sponsorship and examined how the convergence of internal and external threat environment faced by the leader, who is forced to omnibalance, is expected to structure the constraints and incentives for their security behavior, specifically in the realm of providing support for external terrorist groups.

Both the statistical results of the quantitative analysis in Chapter 4 and the qualitative case studies presented in Chapters 5 provided support for the overall argument that state sponsorship is fundamentally a tool of leadership survival. It has been established that neither a conceptualization of state sponsorship as war or state sponsorship as coercive diplomacy can adequately account for the dual external and domestic incentives and constraints that drive leaders to sponsor terrorism. When the basic assumptions of this study: 1) that the state is not a unitary actor; 2) that leaders value political survival above all else; 3) that sponsorship is a rational decision in response to real incentives and constraints; and 4) that leaders establish their regimes in the face of leadership and policy failures of the previous regime are combined with the real world of limited resources, multiple origins of security threats, and the hard trade- offs between economic goods and security, then a framework of leadership survival emerges that can account for the diversity of explanations of state sponsorship in the literature and its utility on the ground. 232 The main hypotheses of the Theory of Leadership Survival, 1) that leaders facing

High External & High Internal (HEHI) threats simultaneously were most likely to sponsor terrorism 2) that as the military gains more resources174 sponsorship becomes more, not less, likely; and 3) that small coalition, autocratic, regimes are most likely to sponsor were all supported by the results of the statistical analysis as well as supported by the historical record in the conforming case studies. The finding that leaders that face an

HEHI threat environment are most likely to sponsor terrorist groups and indications that they are likely to sponsor them for a longer period of time than leaders in the other threat environment categories has major implications for security policy as well as our understanding of how leaders respond to threats beyond the “appease or fight” paradigm.

The HEHI threat environment is particularly important for formulating policy response to state sponsorship because the leader gets caught up in a “strategic bind” (Byman 2005) where their ability to maintain power becomes dependent upon their support for terrorism.

Even the outlier cases of Abdul Razak’s Malaysia and Fujimori’s Peru demonstrated how the underlying dynamics of response to threat were at work, and both helped to flesh out some of the strategic calculations made by leaders in order to remain in power. These cases also helped demonstrate that these dynamics of calculating what caused the threat to their power and formulating effective response were at work for the leaders, whether or not external terrorist groups were the logical tools through which to maintain that power. This dynamic tended to hold regardless of region, level of state

174 Measured here as military expenditure as percentage of GDP. 233 strength, or time period. However, autocratic regimes in an HEHI threat environment were especially susceptible to these pressures and tended to sponsor more frequently.

POLICY IMPLICATIONS As this work wraps up in 2013, the US has officially withdrawn from Iraq, but is still deeply involved in the conflict in Afghanistan. One of the major difficulties of withdrawing is the question of what will happen with Taliban forces and the stability of

Afghanistan when US and coalition forces withdraw. Pakistan is still widely known to be providing support to the Afghan Taliban and al-Qaeda affiliated groups within Pakistan, in the FATA region of Pakistan, and along the Afghan border. It seems that regardless of the amount of foreign aid, intelligence, air support, or diplomatic pressure from the

United States, Pakistan cannot be persuaded to cease support to these groups, many of which are terrorist.

This theory provides a starting point for disentangling some of the key security threats and policy obstacles facing the US and other global powers today. Importantly, the sponsorship situation in Pakistan has shifted from its onset during the time of General

Zia, when it conformed rather closely to the “classic” cases of Libya, Iran, and Iraq outlined in Chapter 5, to a situation where the ISI and the military itself use terrorist groups to further their own foreign policy.175

The migration of this policy from using terrorist groups to protect the state leader to one where it protects the interests of other institutions against the leader, in essence, shifting this protection power from the state leader to the military, is something that

175 There are indications that the dynamics in Iran with the IRGC have shifted in a similar fashion. 234 deserves significant attention in future work, especially considering US policy concerns in the region. While that project is outside the bounds of this study, and will need to be addressed in a future research project, the reframing of state sponsorship as leadership survival allows for some key insights into the problem of Pakistan and a way forward both theoretically and with policy.

First, as long as the leadership of Pakistan experiences internal and external threats simultaneously, leaders will be not only unlikely, but also absolutely unable to cease the aid flowing to the terrorists they have historically supported given that they are trapped in a “strategic bind” (Byman 2005) where the survival of the regime depends on the continued support for terrorism. Second, US policy makers must take the threat reality of the Pakistani leadership into full account and design a policy that ties their cessation of support to the increased external and internal security of the leadership.

Indeed, a key insight from this study is that cessation of support is only likely when the leader experiences more stability and regime protection by ceasing support for terrorists than by continuing. Thus, not only do security concessions from external rivals need to be negotiated, but security guarantees need to be achieved with internal ones as well. In fact, the internal concessions to leader security may be even more vital.

The case of Pakistan, as with the conforming cases in Chapter 5, highlight the fact that policy makers must stop thinking of state sponsorship as something that all leaders can be coerced or enticed out of doing. Threats to strike in response, retaliation, or pre- emption for state sponsorship are likely to increase the HEHI threat environment faced

235 by the leader of the sponsoring state and thus increase, not decrease, sponsorship.176

While HELI threat environment facing leaders can certainly be incentivized and coerced away from support with more ease than HEHI leaders, provided other outlets are available to achieve their foreign policy objectives, HEHI facing leaders are in a completely different situation and it is likely that neither sanctions, nor strikes—short of regime change—can force them to cease sponsorship and may in fact, external coercion to cease sponsorship could have the effect of further entrenching these relationships.

ENDING STATE SUPPORT FOR TERRORISM The first step to halting support for terrorism by state leaders, then, is to understand the incentives and constraints that leaders face which encourage them to sponsor terrorism in the first place. For instance, HELI leaders are likely to act primarily out of strategic interest to achieve foreign policy objectives, normally related to the external threat they face and in conjunction with some policy divergence with their military leadership. Unless this threat is existential, pressure can be put on these regimes to cease support with some success, given that the security of the leader is more closely aligned with the security of the state.

The most effective way to end sponsorship in these situations is with the dissolution, or significant reduction, of the external threat. With resolution of the conflict,

176 Indeed, American airstrikes on Libya had the effect of increasing terrorist attacks against US and Western targets in the years directly following and US threats to invade Iraq had the effect of increasing Saddam Hussein’s support to Hamas, the Palestinian Islamic Jihad, and other groups rather than encouraging the cessation of support even though the US had a history of conventional war with, and strikes against, Iraq. Threats and sanctions against Syria have also not been successful in halting their support for terrorist groups, nor have they been against Iran. In fact, there is every indication that support for terrorism increases as the external threats to the leadership of these countries increase. 236 the strategic objective sought by the regime will either be obtained or decisively out of reach. While it is possible that leaders will continue to pursue covert channels to revive territorial claims or otherwise push against the new status quo, these efforts will be more easily thwarted and sponsorship increasingly unlikely.

For HEHI threat facing leaders, however, the support of terrorist groups quickly becomes intricately connected to the survival of the regime itself, thus certain moves intended to threaten the regime into compliance with counter-terrorism objectives may well have the opposite effect and entrench them into sponsorship more fully. HEHI cases of sponsorship tend to be the most acute, active, and intractable. By conflating these two categories of sponsorship (HELI and HEHI), policy makers not only misdiagnose the problem, but also misunderstand the solution.

Because HEHI leaders are stuck in an extreme “strategic bind” (Byman 2005) where they cannot relinquish terrorism without facing regime collapse, the first step to disentangling the regime from the terrorist group must center on increasing the security of the regime vis-a-vis the internal and external threats they face. The insight that HELI states are still likely to sponsor, but be more amenable to traditional coercion techniques and that LEHI even more less likely to sponsor terrorism may provide some insights into how to go about dismantling of international state-sponsored terrorist infrastructure.

An HEHI state will be most likely to cease sponsorship of terrorism, then, when it’s threat environment can be reduced to one resembling an LEHI threat environment.

While the entrenched nature of sponsorship and very real threat of survival to the regime from the military, even in this situation, may make this threat shift less powerful than if it 237 were not including sponsorship, the mere fact of the reduction of external threat would allow the leader to reorient the military, although many effects and complications of demobilization are still not yet well understood. With the need to mobilize the military for battle removed, the pressure to counter act it will also wane over time. The key is that this will not be an immediate process. It will take time for the leader to downsize and regain full control over the military establishment.

This threat reduction technique can also work by minimizing the internal threat first. If the internal threat posed by the military were dampened, the situation would, over time, transform into an HELI threat environment which, while still more likely to sponsor than LEHI or LELI, has lost the key element of the “strategic bind” and tends to foster more stable civil-military relations (Desch 1999). In this situation, therefore, traditional pressuring techniques such as sanctions, condemnations, threats of strike, etc. would be more likely to be effective at ending the sponsorship and less likely to further entrench the support.

Ultimately, short of external regime change of leaders in HEHI threat environments, which although very effective for ceasing support to terrorists, brings with it an entirely different set of new problems that the world is largely unprepared to manage, this study concludes that the most effective way to end state sponsorship is to find ways to assist leaders in alleviating the most pressing threats to their rule. This naturally presents a policy conundrum for those who have no desire to “coddle dictators,” yet, the success of international security policy requires not only clear objectives, but also realistic, fact based approaches. If reducing or eliminating state sponsored terrorism is the 238 policy objective, a new approach will need to be taken that devises ways to secure the autocratic leader against, and thus reduce, the HEHI threat or will need to decide that regime change is the only politically acceptable way end this type of sponsorship.

Understanding the difference in sponsorship motivations between HEHI and HELI facing leaders is key to developing policy responses to state sponsorship that are truly effective at stopping support for terrorism.

PATHWAYS FOR FUTURE RESEARCH This study has opened numerous pathways for future research, some case-specific and others more general to the topic. Also, while this research did not explicitly touch on the issue of weak or failing states as havens for terrorism, there is the possibility that the basic logic of leadership survival is at work here as well, although the exact dynamics may be different than in stronger HEHI states. This section will layout pathways for future research suggested by this study.

Case specific pathways There are several cases that deserve a closer empirical look. For example, sponsorship patterns during Iraq’s third HEHI period—the run up to the Iraq war—once the dataset is updated should be examined to see if a similar pattern emerges to the first two periods examined in Chapter 5. Indeed, process tracing the relationship between the leader and the military in all cases of sponsorship would further shed some light on this dynamic. Closer looks into the military-leader-terrorist group relationships in cases such as Syria under both Hafez and Bashar al-Assad, in Azerbaijan under Alyiev, Pakistan

239 under Zia al-Huq, and Sudan under al-Bashir could provide further refinement of this theory.

Cases of sponsorship in sub-Saharan Africa specifically remain under studied.

While this project did make a point to explicitly include them in the quantitative analysis, the high levels of HEHI threat environment faced by some African leaders, even today, provide an additional crucial post-Cold War test of this theory. The conceptualization of both terrorism and state sponsorship provided by this study allows scholars to continue to step away from the big and spectacular incidences of terrorism that have defined our understanding and toward understanding terrorism as something that can occur small- scale and in contexts outside those that are likely to receive the most attention from

Western media.

While some pioneering studies into the nature of the state-terrorist relationship have been done by Daniel Byman and Sarah Kreps (2010) there remain many unanswered questions. For example, it is still unclear as to how the institutionalization of state sponsorship occurs at the state level. In addition, there are indications that the motivations that drive a leader to provide support to a terrorist group initially may not be the same as the motivations that encourage a leader to continue that support and the nature and implications of that path dependency should be further explored.

Furthermore, there appear to be shifts in the locus of power within the state because of state sponsorship, such as has occurred in Pakistan, a situation that threatens to be reproduced with the IRGC in Iran. There are questions as to whether this dynamic is best understood as a “hijacking” of foreign policy by a domestic faction or whether it is a 240 type of power entrenchment that functions to balance rival actors in the decision-making apparatuses of the state. It is possible that this is what “mature” state sponsorship looks like. If that is the case, bringing the long term effects of state sponsorship to the attention of potential sponsors, especially leaders facing HEHI threat environments, and could encourage them to take a different path and thus prevent the sponsorship from occurring in the first place.

While some of these cases, such as Syria and Pakistan, have been examined in depth before, revisiting them using a framework provided by the Theory of Leadership

Survival promises to shed new light on the dynamics of state sponsorship and to provide some clear direction as to how to develop and implement clear and effective policies to counter it. This study has also demonstrated that while HEHI leaders tend to become entrenched in sponsorship, HELI leaders are likely to use it as a policy tool as well.

HELI Threat Environments and Sponsorship This study found empirical support for the notion that leaders facing an external threat, across the board, were likely to sponsor terrorist groups. However, it had much less leverage in cases where the leader faced only an external threat (HELI), possibly because of the rough calibration of the internal threat variable, which was only able to account for extreme preference divergence between the leadership and the military in the form of threat of coup d’état. Further research should be done in order to examine how preference divergence incentivizes a leader to sponsor in the HELI threat environment.

Like in the HEHI cases, there are indications that the shadow of the previous leader’s policy failures plays a tremendous role in HELI instances of sponsorship. 241 Territorial conflicts or other issues that led to the defeat of the previous ruler become the issues that the legitimacy of the next ruler becomes staked upon. If the leader depends on the achievement of the policy in order to maintain their rule and the military refuses to comply, terrorist groups or other non-state actors can be used to circumvent the military, achieve the goal, and maintain the legitimacy of their rule. Additional work should delve deeper into the nuances of civil-military relations in HELI cases and develop a sensitive metric for security policy diversion between the state leader and the military establishment.

Effect of State Sponsorship on International Relations This study has also raised questions about how state sponsorship affects international relations. For instance, what are the long-term effects of state sponsorship on the international state system and on the relationship between states within it? Areas of study such as the effects state sponsorship on the level of conflict between states, the impact on foreign policy development, grand strategy, and international law have not received adequate attention by scholars. Especially, it will be useful to examine how exactly state sponsorship has shaped the calculations of risk between states, such as Iran and Israel.

One particular area that has been under examined, with significant policy and security implications, is the question of whether state sponsorship can work to create stable deterrence regimes between states, or whether the unpredictability of the non-state actors and the dynamic relationship between the terrorists and the state inevitably result in instability and escalation. 242 In short, the grand strategy of the United States was completely reconfigured because of state sponsored terrorism, multiple international wars have been fought,177 and numerous international laws and conventions implemented, sanctions and strikes authorized, and regimes overthrown all without a good understanding of the causes and effects of state sponsorship in general, let alone the specific policies. This study provides a starting point for solid and broad empirical study of state sponsorship as a phenomenon linked to leadership survival, but there is still a tremendous amount left to learn.

CONCLUSION While additional research and refinement of the theory are obviously necessary, this study has made a tremendous empirical and theoretical contribution to how we understand the political logic of state sponsored terrorism. It has not only brought the state leaders in as actors, but has highlighted that leaders have markedly different interests than the state. Not only are these interests different, when faced with threats to their rule leaders will often make choices to secure their own positions of power even at the expense of the state. The incentives and constraints faced by these leaders are the key level of analysis to examine the decision to sponsor terrorism.

This study has also introduced the military as a key factor in the study of state sponsorship, one that has been surprisingly ignored in previous works. It has also examined the role that security services play as a link between the leader and the terrorist groups and the role that this plays in helping to balance against the military and internal threats of coup d’état to the leader. Most importantly, it has shown how the intersection

177 The 1982 invasion of Lebanon, the Iraq War and war in Afghanistan, just to name a few. 243 of high external and internal threats to the regime create the situation where leaders are likely to choose a policy of state sponsorship in order to balance these threats.

This study has contributed to multiple research agendas that deal with civil- military relations, internal war, interstate conflict, terrorism, regime stability, and leader survival. As we move further into the 21st century, where interstate war is less traditionally conventional and more confined to cross-border skirmishes, where most militaries, especially in the developing world, still rely on a gun rather than the rapidly developing military technologies of the United States and industrialized west, and non state actors are more prevalent on the world stage, pursuing their own foreign policies, developing economic and political connections with other groups, as well as with states, we will see dramatic shifts in how war is fought, by whom, and with what material.

Terrorism, however, is as old as politics, and this study has indicated that leaders facing

HEHI threat environments will continue to be the most likely to pursue relationships with terrorist groups in the attempt to secure their positions of power.

By refocusing counter-terrorism policy, then on the real threats leaders face and developing programs that work to assist state leaders in reducing the level of these threats, the strength of terrorist groups can be minimized as the resources, and other assets provided by the nation-state are cut off, significantly decreasing their capabilities and operational strength.

The utility of this empirical foundation for counter-terrorism cannot be overstated.

Moving forward, it increases our predictive capabilities, opens new pathways for prevention and effective policy making against sponsors, reduces the need for 244 conventional military forces to battle state-backed terrorist groups, and will provide leaders with real alternatives to terrorist groups in order to preserve their regimes.

245 Appendix A

State Terrorist Group Sponsored? State Sponsor October 80 NO September 11 NO "1 May" NO 15 May Organization YES Iraq 1st of May Group NO 20 December Movement NO 23rd of September Communist League NO 2nd of June Movement NO Abu al-Rish Brigades NO Sudan, Libya, Syria, Iraq and Iran, North Abu Nidal Organization178 YES Korea, and China Abu Sayyaf Group YES Iran, Sudan and Libya Achik National Volunteer Council YES Bangladesh, Bhutan, Burma, and Pakistan Action Directe NO Aden Abyan Islamic Army NO Mozambique, Cuba, Soviet Union, and African National Congress YES Eastern bloc Afrikaner Resistance Movement (AWB)179 NO Al Jehad NO Al Sadr Brigades NO Al Zulfikar YES Afghanistan and India Al-Adl Wal Ihsane NO Al-Aqsa Martyrs Brigades YES Iran Al-Da'wah Party YES Iran Soviet Union and East European states, Jordan, Syria, Saudi Arabia, Kuwait, China Al-Fatah YES and North Korea. Al-Fatah Uprising YES Syria Al-Gama'a al-Islamiyya (GAI) YES Sudan and Iran Somalia, Sudan and Kenya (defacto), and Al-Ittihaad al-Islami (AIAI) YES Eritrea.

178 Aka Black June; Fatah Revolutionary Council; Arab Revolutionary Brigades, Black September; Revolutionary Organization of Socialist Muslims. 179 Aka Boer Attack Troops, Afrikaaner Weestand Beweeging

229 Afghanistan, Yemen, Somalia, Sudan and Al-Qaeda YES Niger Al-Saiqa YES Syria Al-Umar Mujahedeen YES Pakistan Albanian National Army (ANA) NO Alejo Calatayu NO Alex Boncayao Brigade (ABB NO Alfaro Vive Carajo (AVC) YES Nicaragua, Cuba, and Libya All Burma Students' Democratic Front (ABSDF) NO All India Anna Dravida Munetra Kazgan Party NO All India Sikh Students Federation (AISSF) NO All Tripura Tiger Force (ATTF) YES Bangladesh and Pakistan Allah's Tigers NO Allied Democratic Forces (ADF) YES Sudan and Zaire Amal YES Syria Anarchist Struggle NO Animal Liberation Front (ALF) NO Animal Rights Militia NO Anti-Castro Command NO Anti-Imperialist International Brigade NO Anti-State Action NO Anti-terrorism ETA (ATE) NO Anti-Terrorist Liberation Group NO April 19 Movement (M-19) YES Cuba, Nicaragua, and Libya Arab Communist Organization (ACO) NO Arab Liberation Front (ALF) YES Iraq Arab Nationalist Youth for the Liberation of Palestine (ANYLP) YES Created by Qaddafi in 1974 Argentine Anti-Communist Alliance aka Triple A NO Argentine Liberation Front (FAL) NO

230 Argentine National Organization Movement (MANO) NO Armata Corsa NO Armed Commandos of Liberation NO Armed Communist Struggle NO Armed Forces of National Liberation (FALN) MAYBE Possibly Cuba Armed Forces of National Resistance (FARN) YES Cuba Armed Forces of Popular Resistance (FARP) NO Armed Forces Revolutionary Council (AFRC) NO Armed Islamic Group (GIA) YES Sudan and Iran Armed Nucleus for Popular Autonomy NO Armed Proletarian Nuclei (NAP) NO Armed Proletarian Squads NO Armed Revolutionary Nuclei (NAR) NO Armed Squads for Communism NO Armenian Secret Army for the Liberation of Armenia (ASALA) YES Soviet Union Libya, Syria, and Iran Army of God NO Aryan Nations (AN) NO Asbat al-Ansar NO Aum Shinrikyo / Aleph NO Babbar Khalsa International (BKI) YES Pakistan Libya, Algeria, Cuba, Nicaragua, South Basque Fatherland and Yemen, Algeria, Lebanon, and the Soviet Freedom (ETA) YES Union. Black Liberation Army YES Cuba Black Panthers YES Cuba, Algeria and North Korea. Black Panthers (West Bank/Gaza) NO Black September Organization YES Libya Black War NO

231 Bodo Liberation Tigers (BLT) YES Pakistan Breton Revolutionary Army (ARB) NO Canary Islands Independence Movement NO Soviet KGB, East German Stassi Carlos the Jackal YES (Anderson), Cuba (CIA GPoT 1976). Catholic Reaction Force (CRF) NO Charles Martel Group NO Chukakuha aka Middle Core Faction, Nucleus Faction NO Cinchoneros Popular Liberation Movement aka Cinchonero People's Liberation Movement YES Cuba and Nicaragua. Commando Internacionalista Simon Bolivar NO Maybe Libya, but unlikely Committee for the Security of the Highways NO Communist Combatant Cells NO Communist Party of Nepal- Maoist (CPN-M) NO Continuity Irish Republican Army (CIRA) NO Contras180 YES United States and Israel Croatian Freedom Fighters (CFF) NO Death to Bazuqueros NO Death to Kidnappers NO Delta Group NO Democratic Front for Renewal NO Democratic Front for the Liberation of Palestine Syria, Libya, USSR, Cuba, and South (DFLP) YES Yemen Dev Sol YES Greece and Syria Dev Yol NO DHKP/C (splinter of Devrimci Sol (Dev Sol)) NO

180 Aka Counter Revolutionaries (primarily under the United Nicaraguan Opposition) but the Nicaraguan Democratic Force was another early group formed in 1981. Also the Democratic Revolutionary Alliance (ARDE).

232 Dignity for Colombia NO Dima Halam Daoga (DHD) YES Pakistan Direct Action Against Drugs (DADD) NO Eagles of the Palestinian Revolution YES Syria Earth Liberation Front NO East Asia Anti Japanese Armed Front aka Higashi Ajia Hannichi Buso Sensen NO East Turkistan Liberation Organization NO Egypt's Revolution NO Egyptian Islamic Jihad (EIJ) YES Iran and Sudan Eritrean Liberation Front Sudan, Syria, Iraq, the USSR, the UAE, and (ELF) YES Saudi Arabia Eritrean People's Liberation Front (EPLF) YES Libya, Syria, Iraq, Sudan Ethiopian People's Revolutionary Army (EPRA)181 YES Sudan Evan Mecham Eco-Terrorist International Conspiracy (EMETIC) NO Farabundo Marti National Cuba, Nicaragua, Soviet Union, Vietnam Liberation Front (FMLN) YES and possibly some Middle Eastern states. Fatah Hawks NO February 28 Popular Leagues NO Fedayeen Khalq (People's Commandos) NO Federation of Students and Scholars of Cote d'Ivoire (FESCI) NO Fighting Guerrilla Formation NO First of October Antifascist Resistance Group (GRAPO) NO Force 17 NO Forces of the Struggling Ranks NO Forqan Group NO Forum for the Restoration of Democracy-Kenya NO

181 Was the armed wing of the Ethiopian People's Revolutionary Party (EPRP)

233 Francisco Villa People's Front NO Francs Tireurs (Mavericks) NO Free Aceh Movement (GAM) MAYBE Alleged Libyan support Free Fatherland aka Partido Patria Libre NO Free Galician People's Guerrilla Army NO Free Papua Movement (OPM) YES Libya Free South Moluccan Youth's NO Freedom Party NO Front for the Liberation of Lebanon from Foreigners (FLLF) NO Front For the Liberation of the Azores NO Front for the Liberation of the Cabinda Enclave (FLEC- FAC/FLEC-R) YES DRC, France, Zaire, and Gabon Front for the Liberation of the French Somali Coast NO Fronte di Liberazione Naziunale di a Corsica (FLNC) MAYBE Unconfirmed reports of links with Iran Fuerza Nueva (New Force) (FN) NO GAC NO George Jackson Brigade NO Gracchus Babeuf NO Great Eastern Islamic Raiders Front (IBDA-C) NO Greek National Socialist Organization NO Grey Wolves NO Perhaps Turkic Central Asian states Group of the Martyrs Mostafa Sadeki and Ali Zadeh NO Grupo Estrella NO Gruppe Haw Weg Den Scheiss NO Guadeloupe Liberation Army NO Guardsmen of Islam MAYBE Probably Iran

234 Guatemalan Communist Party NO Guatemalan Labor Party NO Guatemalan National Revolutionary Unity (URNG) YES Cuba, Mexico, and Nicaragua Guerrilla Army of the Poor YES Cuba Guerrillas of Christ the King NO Gurkha National Liberation Front (GNLF) NO Hamas YES Jordan, Syria and Sudan, and Iran. Harakat ul-Jihad-i-Islami (HUJI-B) YES Bangladesh and Pakistan Harakat ul-Jihad-i-Islami (HUJI) YES Afghanistan Harakat ul-Mujahidin (HuM) YES Pakistan, maybe Afghanistan Pakistan, perhaps some sympathetic Arab Harkat ul-Ansar YES countries Hector Riobe Brigade NO Hezbollah YES Iran, Sudan, and Syria Hizb-I-Islami YES United States and Pakistan Hizbul Mujahideen (HM) YES Pakistan Holy Spirit Movement NO Iberian Liberation Movement (MIL) NO Idealist Association aka Yusufeli Ülkücü NO Ikhwan-ul-Muslimeen YES Afghanistan Independent Armed Revolutionary Commandos (CRIA) YES Cuba Independent Armed Revolutionary Movement (MIRA) YES Cuba Independent Nasserite Movement NO Inkatha Freedom Party (IFP) NO International Revolutionary Action Group (GARI) NO International Revolutionary Solidarity NO Iparretarrak (IK) MAYBE France (government complicity unclear) Iraqi Islamic Vanguards for National Salvation (IIVNS) YES Iran

235 Iraqi Liberation Army NO Iraqi Mujahedeen MAYBE Perhaps Iran Irish National Liberation Army (INLA) NO Irish People's Liberation Organization (IPLO) NO Irish Republican Army (IRA) YES Libya, Iran, Algeria, possibly Sudan. Islam Liberation Front NO Islami Chhatra Shibir (ICS) YES Pakistan Islami Inqilabi Mahaz NO Islamic Front for the Liberation of Palestine (IFLP) NO Islamic Great Eastern Raiders Front (known by the acronym IBDA-C) NO Islamic Jihad Organization (Yemen) NO Islamic Liberation Organization NO Islamic Movement of Uzbekistan (IMU) YES Pakistan and Saudi Arabia Islamic Salvation Front (FIS) aka Armee Islamique du Salut YES Sudan and Iran Italian Social Movement (MSI) NO Ittehad-i-Islami YES Saudi Arabia Jagrata Muslim Janata Bangladesh182 NO Jaime Bateman Cayon Group (JBC) NO Jaish-e-Mohammad (JeM) YES Pakistan and Afghanistan Jamaat-al-Fuqra NO Jamaat-E-Islami (Bangladesh) MAYBE Maybe Pakistan Jamaat-E-Islami (India/Pakistan) YES Pakistan Jamaica Labor Party NO Jamiat-e Islami-yi Iran, Russia, Uzbekistan, India, and Afghanistan YES Tajikistan. Jammu and Kashmir Islamic YES Pakistan

182 Aka Shanti Raksha Committee (Peace-Keeping Committee)

236 Front (JKLF) Janashakti NO January 22 group NO January 31 Popular Front NO Japanese Red Army (JRA)183 YES North Korea, Syria, Lebanon, Libya. Jemaah Islamiya (JI) aka Conservative Persian Gulf states such as Islamic Group; Islamic Saudi Arabia, government complicity Community MAYBE unclear. Jewish Armed Resistance NO Jewish Defense League (JDL) NO Jordanian Islamic Resistance NO Jorge Eliecer Gaitan Nationalist Movement NO June 16 Organization NO Justice Commandos for the Armenian Genocide NO Kach aka Kahane Chai NO Kachin Independence Army (KIA)184 YES China and India Kanak Socialist National Liberation Front NO Kanglei Yawol Kanna Lup (KYKL) NO Karenni National Progressive Party185 NO Kashmiri Hizballah NO Kayin National Union (KNU) aka Karen National Union YES India and possibly Thailand Kenya African National Union (KANU) NO Keshet NO Khalistan Commando Force NO Khalistan Liberation Force (Khalistan Liberation Front) YES Pakistan Khmer Rouge YES China and possibly Thailand Khristos Kasimis Revolutionary Group for International Solidarity NO

183 Aka Anti-Imperialist International Brigade 184 The armed wing of the Kachin Independence Organization (KIO) 185 Political wing of the Karenni Army

237 Kosovo Liberation Army (KLA) note: in Albanian it is UCK YES Albania, Macedonia, some NATO countries Ku Klux Klan (KKK) NO Kurdish Democratic Party (KDP) YES Iraq, Iran, and Syria Syria, Greece, the Soviet Union, Kurdistan Workers' Party Czechoslovakia, South Africa, Armenia and (PKK) YES Italy. Landless Peasants' Movement (MST) NO Lashkar-e-Jhangvi (LeJ) YES Afghanistan Lashkar-e-Omar (LeO) NO Lashkar-e-Taiba (LeT) YES Pakistan Afghanistan and elements of the Indonesian Laskar Jihad YES military Lautaro Youth Movement YES Cuba Lebanese Armed Revolutionary Faction NO Lebanese Liberation Front NO Lebanese National Resistance Front NO Lebanese Socialist Revolutionary Organization NO Lesotho Liberation Army (LLA) NO Liberation Tigers of Tamil Eelam (LTTE) aka Tamil Tigers YES India Lord's Resistance Army (LRA) YES Sudan Lorenzo Zelaya Revolutionary Front (LZRF) NO Loyalist Volunteer Force (LVF) NO Macheteros NO Mano Blanca (White Hand) NO Manuel Rodriguez Patriotic Front (FMR) YES Cuba, other communist states Maoist Communist Center (MCC) aka Naxalites YES Nepal and Pakistan Martyr Abu Ja'far Group NO Marxist-Leninist Armed Propaganda Unit NO

238 Masada, Action and Defense Movement NO Maximiliano Gomez Revolutionary Brigade NO Maximiliano Hernandez Martinez Brigade NO May 19 Communist Order NO Meibion Glyndwr NO Miskito Indian Organization NO Mizo National Front NO Mohajir National Movement (MQM) aka Muhajir Quami Mahaz YES UK safe haven Mon Guerrillas NO Mong Thai Army (MTA) NO Montoneros YES Cuba Morazanist Patriotic Front (FPM) YES Cuba and Nicaragua Moro Islamic Liberation Front (MILF) YES Iran and Libya Moro National Liberation Front (MNLF) YES Libya, Malaysia, and Iran Movement for Actualization of the Sovereign State of Biafra (MASSOB) NO Movement of Democratic Guinea -Bissau Mauritania, Gambia, Libya, Forces of Casamance YES and Iraq. Movement of the Revolutionary Left aka Movimiento de la Izquierda Cuba, Angola, Algeria, Mozambique, and Revolucionaria (MIR) YES Nicaragua. Mozambique National Resistance Movement (RENAMO) previously known as (MNR) YES Rhodesia and South Africa Mujahedin-e-Khalq (MeK) YES Iraq, safe haven in France Mujahedeen KOMPAK NO Muslim Brotherhood YES Jordan and Iraq Muttahida Qami Movement (MQM) NO Nation of Yahweh NO National Anti Communist Commando NO DRC, no indication of govt. complicity National Army for the YES Sudan

239 Liberation of Uganda (NALU) National Committee Against Independence (CNCI) NO National Council for the Defense of Democracy186 YES DRC, Zimbabwe, Zaire and Tanzania National Democratic Front of Bodoland (NDFB) YES Bangladesh, Myanmar and Bhutan National Front Against Tigers (NFAT) NO National Front for the Liberation of Cuba (FLNC) NO National Integration Front (FIN) NO National Liberation Army (Colombia) (ELN) YES Cuba and Venezuela National Liberation Army of Bolivia (ELN) YES Cuba and the Soviet Union National Liberation Front of Tripura (NLFT) YES Bangladesh and Pakistan National Organization of Cypriot Fighters (EOKA) NO National Patriotic Front of Liberia (NPFL) YES Burkina Faso, Libya, and Cote d'Ivoire National Resistance Movement (RNM) NO Libya until 1982 National Socialist Council of Nagaland YES China National Socialist Council of Nagaland-Isak-Muivah (NSCN-IM) YES Bangladesh National Socialist Council of Nagaland-Isak-Muivah (NSCN-IM) YES China and alleged links to Pakistan National Union for the Total DRC, South Africa, the United States, Independence of Angola France, Israel, Corte d'Iviore, Burkina Faso, (UNITA) YES and Togo. Nestor Paz Zamora Commission NO New Armenian Resistance (NAR) NO New Order NO New Order-France NO

186 Also National Council for the Defense of Democracy–Forces for the Defense of Democracy (CNDD- FDD)

240 New Patriotic Party (NPP) NO New Pattani United Liberation Organization (New PULO) NO New People's Army (NPA)187 YES China and Soviet Union New Red Brigades/Communist Combatant Party NO New World Liberation Front (NWLF) NO Nicaraguan Resistance (Contra) YES United States Ninjas YES Zaire Odua Peoples' Congress NO Official IRA NO Omar Torrijos Commando for Latin American Dignity NO Omega-7 NO Orange Volunteers (OV) NO Organization of the Sons of Occupied Territories NO Organization of Volunteers for the Puerto Rican Revolution NO Organized Comrades for Feminist Counter-Power NO Organized Proletarian Communists NO Orly Organization NO Oromo Liberation Front (OLF) YES Eritrea, Somalia, Sudan Pakistani People's Party (PPP) NO Palestine Liberation Front YES Libya, Syria, and Iraq Palestine Liberation Organization (PLO) YES Libya and the Soviet Union. Palestinian Islamic Jihad (PIJ) YES Syria and Iran Palestinian Popular Struggle Front (PSF) YES Jordan, Lebanon and Syria Pan Africanist Congress (PAC) YES China and Tanzania

187 The armed wing of the Communist Party of the Philippines.

241 Party for the Liberation of the Hutu People (PLPH) aka Palipehutu-FNL YES Rwanda, Tanzania, and the DRC Patria Nueva (New Country) NO Patriotic Union of Kurdistan (PUK) YES The United States, Iran, Syria, and Libya Pattani United Liberation Organization (PULO) YES Libya Peasant Self-Defense Group (ACCU) NO Pedro Albizu Campos Revolutionary Forces NO Pedro Leon Arboleda Movement NO People Against Gangsterism And Drugs (PAGAD) YES One faction, Qibla, is backed by Libya. People's Alliance NO People's Liberation Army Bangladesh and Pakistan, defacto support (PLA) YES of Myanmar People's Liberation Forces aka Farabundo Marti Popular Liberation Forces NO People's Liberation Front (JVP) NO People's Liberation Organization of Tamil Eelam MAYBE India, complicity unknown People's Revolutionary Army (Argentina) aka Ejercito Revolucionario del Pueblo (ERP) YES Cuba People's Revolutionary Army (Colombia)188 NO People's Revolutionary Army (ERP) (El Salvador) NO People's Revolutionary Command (CRP) NO People's Revolutionary Organization NO People's Revolutionary Organization- Colombia (ORP) NO People's War Group (PWG)189 YES Nepal and Pakistan

188 Aka Ejercito Revolucionario del Pueblo (ERP)

242 Peronist Armed Forces (FAP) NO Saudi Arabia, Kuwait, Syria, the United Phallange YES States, and Israel. Algeria, Angola, Botswana, China, Cuba, Libya, Mozambique, USSR, Tanzania, Polisario Front YES Zambia. Poor People's Party NO Popular Forces of April 25 NO Popular Front for the Liberation of Palestine -- General Command (PFLP- Syria, Iraq, Iran, Libya, Soviet Union and GC) YES communist Eastern Europe. Popular Front for the Liberation of Palestine (PFLP) YES Syria, Libya, the Soviet Union, and China. Popular Liberation Army aka Ejército Popular de Liberación (EPL) NO Popular Militia (Colombia) NO Popular Resistance Committees NO Popular Revolutionary Army NO Popular Revolutionary Bloc (BPR) NO Popular Revolutionary Resistance Group NO Popular Revolutionary Vanguard NO Popular Self-Defense Forces (FAP) aka Mai-Mai YES DRC Prima Linea NO Proletarian Action Group NO Proletarian Patrols NO Protestant Action Group NO Puerto Rican Armed Resistance 190 NO Purbo Banglar Communist Party (PBCP) NO Quintin Lame NO Ranbir Sena NO

189 Aka Naxalites, People’s Guerrilla Army (PGA), People's War (PW), The Communist Party of India- Marxist Leninist (People's War) CPI-ML (PW) 190 Aka Puerto Rican Resistance Movement

243 Real Irish Republican Army (RIRA) NO Rebel Armed Forces (FAR) (of Guatemala) YES Cuba Recompras NO Recontras NO Red Army Faction aka German Democratic Republic, Baader Meinhof YES Czechoslovakia, South Yemen, and Libya. Red Brigades YES Libya and France Red Brigades Fighting Communist Party (BR-PCC) NO Red Cell NO Red Flag (Venezuela) aka Bandera Roja YES Has had links with the Cuban government Red Flying Squad NO Red Guerilla Family NO Red Guerrilla NO Red Hand Commandos NO Red Hand Defenders (RHD) NO Red Revolutionary Front NO Reform of the Armed Forces Movement NO Republic of New Afrika NO Republican Action Force NO Resistenza NO Revolutionary Anti-Racist Action NO Revolutionary Armed Forces of Colombia (FARC) YES Venezuela, Soviet Union, and Cuba. Revolutionary Cells (RZ) aka Revolutionaere Zellen NO Revolutionary Commandos of the People (CRP) YES Nicaragua and Cuba Revolutionary Communist League (LCR) (Spain) NO Revolutionary Eelam Organization (EROS) NO Revolutionary Front for an Independent East Timor (FRETILIN) NO Revolutionary Nuclei NO Revolutionary Organization 17 November NO Revolutionary Organization NO

244 of People in Arms (ORPA) Revolutionary Organization of Socialist Moslems YES Iraq, Syria, and Libya Revolutionary Patriotic Anti- Fascist Front (FRAP) NO Revolutionary People's Struggle (ELA) aka Revolutionary Popular Struggle NO Revolutionary Perspective NO Revolutionary Popular Left NO Revolutionary Proletarian Initiative Nuclei NO Revolutionary Solidarity NO Revolutionary United Front Liberia, Libya, Burkina Faso and Cote (RUF) YES d'Ivoire Revolutionary Vanguard NO Revolutionary Worker Clandestine Union of the People Party NO Revolutionary Workers Party NO Revolutionary Workers Party of Bolivia (PRTB) NO Revolutionary Workers' Council (Kakurokyo) NO Ricardo Franco Front (RFF) NO Robin Food NO Roderigo Franco Command NO Rote Zora NO Rwanda Patriotic Front YES Uganda Salafist Group for Preaching and Fighting (GSPC) NO Sanidila Secessionist Movement NO Saor Eire (Irish Republican Group) NO Scottish National Liberation Army NO Secret Anti-Communist Army (ESA) NO Secret Cuban Government NO Secret Organization Zero NO Seikijuku NO Sekihotai NO

245 Senki ("Battle Flag") NO September 11 Commandos NO Shanti Bahini - Peace Force NO Shining Path (Sendero Luminoso) (SL) NO Simon Bolivar Guerilla Coordinating Board (CGSB) NO Sipah-e-Sahaba/Pakistan (SSP) NO Sipah-I-Mohammed NO Social Resistance NO Socialist Patients' Collective (SPK) NO Soldiers of Truth NO Somali National Movement YES Ethiopia South-West Africa People's Organization (SWAPO) YES Libya, Soviet Union, Angola (dataset) Southern Sierra Peasant Organization NO Spanish Armed Group NO Spanish Basque Battalion NO Students Islamic Movement of India (SIMI) YES Pakistan Sudan People's Liberation Egypt, Eritrea, Ethiopia, Israel, Libya, Army (SPLA) was the armed Uganda, Saudi Arabia, Zaire, indirectly by wing of the SPLM YES the USA Supreme Council for Islamic Revolution in Iraq (SCIRI) YES Iran, Syria, UKG and Kuwait Survivors of Golfech NO Sword of Islam NO Symbionese Liberation Army (SLA) NO Syrian Social Nationalist Party YES Syria Takfir wa Hijra NO Taliban YES Paki stan Tamil Eelam Liberation Organization (TELO) YES India Tanzim NO Tehrik al-Mojahedin; Tehreek-ul-Mujahedeen (TuM) YES Pakistan Terra Lliure (TL) NO Terror Against Terror NO

246 The Extraditables NO The Order (Silent Brotherhood) NO Tigray Peoples Liberation Front (TPLF) YES Sudan, Syria, and Somalia TKP/ML-TIKKO191 NO Tribal Battlefront NO Tripura National Volunteers (TNV) YES Bangladesh Tuareg Guerrillas NO Tupac Amaru Revolutionary Movement YES Cuba, Nicaragua, Libya Tupac Katari Guerrilla Army (EGTK) NO Tupamaro Revolutionary Movement - January 23 NO Tupamaros (Movimento de Liberacion National, MLN) YES Cuba Turkish Communist Party/Marxist (TKP-ML) NO Turkish Communist Workers Party NO Turkish Hezbollah MAYBE Iran suspected Turkish People's Liberation Army (TPLA) YES Soviet bloc countries Turkish People's Liberation Front YES Soviet bloc countries Uganda Freedom Movement (UFM) YES Libya Uganda People's Army NO Tanzania in 1972 Uighur Separatists YES Afghanistan Ulster Defence Association/Ulster Freedom Fighters NO Ulster Volunteer Force (UVF) NO Union Guerrera Blanca (UGB) NO Union of the People (UDP) NO United Democratic Front (UDF) NO United Freedom Front (UFF) NO

191 Aka Maoist Communist Party, Maoist Komunist Partisi, People's Liberation Army, Turkish Communist Party/Marxist Leninist, Turkish Workers' and Peasants' Liberation Army

247 United Liberation Front of Pakistan, Bangladesh, possibly China and Assam (ULFA) YES Cambodia United Liberation Movement for Democracy in Liberia (ULIMO) NO United Nasserite Organization YES Libya United National Party NO United People's Democratic Solidarity (UPDS) YES Bhutan and Bangladesh. United Popular Action Front (FAPU) NO United Popular Action Movement NO United Self-Defense Forces of Colombia (AUC) NO Up the IRS, INC. NO Vietnamese Party to Exterminate the Communists and Restore the Nation NO Vigorous Burmese Student Warriors NO Vishwa Hindu Parishad (VHP) NO Vitalunismo NO Waltraud Boock Group NO Weather Underground aka Weathermen NO West Nile Bank Front (WNBF) YES Sudan and the DRC White Legion (Georgia) NO White Wolves NO Workers' Organization for Communism NO Workers' Self-Defense Movement (MAO) NO Youth Action Group NO Zapatista National Liberation Army (ELZN) NO Zarate Willka Armed Forces of Liberation NO Zebra killers NO Zimbabwe African Nationalist Union (ZANU) YES China Zimbabwe African People's YES Mozambique, Zambia, Tanzania, the Soviet

248 Union (ZAPU) Union and Cuba Zimbabwe Patriotic Front NO Zimbabwe People's Army (ZIPA) YES Mozambique

249 Bibliography

Aburish, Said K., Saddam Hussein: The Politics of Revenge, New York, NY: Bloomsbury Publishing, 2000.

Adams, Brad, “India’s shoot-to-kill policy on the Bangladesh border,” The Guardian UK January 23, 2011, available online at http://www.guardian.co.uk/commentisfree/libertycentral/2011/jan/23/india- bangladesh-border-shoot-to-kill-policy accessed October 21, 2012.

Adams, James. "The financing of terror: how the groups that are terrorizing the world get the money to do it" New York, NY: Simon and Schuster, 1986.

Addison, Tony & S. Mansoob Murshed. “Transnational terrorism as a spillover of Domestic disputes in other countries,” Defense & Peace Economics. Vol. 16, No. 2 (April 2005) pp. 69-82.

Al-Marashi, Ibrahim and Sammy Salama, Iraq’s Armed Forces: An Analytical History. New York, NY: Routledge, 2008.

Alagappa, Muthiah, “Asian Military Relations: Key Developments, Explanations, and Trajectories.” In Coercion and Governance: The Declining Political Role of the Military in Asia. Muthiah Alagappa, ed. Stanford, CA: Stanford University Press, 2001.

Allison, Graham, The Essence of Decision: Explaining the Cuban Missile Crisis. Boston, MA: Little, Brown and Company, 1971.

Anderson, Sean K. & Stephen Sloan, Historical Dictionary of Terrorism. 2nd edition. Lanham, Maryland: Scarecrow Press, 2002.

Andrew, Christopher and Vasili Moitrokhin, "The World was going our way: the KGB and the Battle for the third world." New York, NY: Basic Books, 2005.

Arendt, Hannah, The Origins of Totalitarianism. New York, NY: Harcourt, Brace, 1951.

Ariel L. Ahram, Proxy Warriors: The Rise and Fall of State-Sponsored Militias,” Stanford, CA: Stanford University Press, 2011.

Arnold, Guy, The Maverick State: Gaddafi and the New World Order, New York, NY: Cassell Wellington House, 1996.

Bapat, Navin, “State Bargaining with Transnational Terrorist Groups” International Studies Quarterly, Vol. 50 No. 2 (2006): 215-232.

250 ______. “Understanding State Sponsorship of Militant Groups.” British Journal of Political Science Vol. 42, No.1 (January 2012) pp. 1-29.

______. “The Internationalization of Terrorist Campaigns.” Conflict Management and Peace Science. Vol. 24 No. 2 (2007): 265-280.

Blundy, David & Andrew Lycett, Qaddafi and the Libyan Revolution, Boston, MA: Little, Brown and Company, 1987.

Brooks, Risa A. Shaping Strategy: The Civil-Military Politics of Strategic Assessment. Princeton, NJ: Princeton University Press, 2008.

Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow, The Logic of Political Survival, Cambridge, MA: MIT Press, 2003.

Bush, George W. 2002 State of the Union Address. Available online at http://stateoftheunionaddress.org/2002-george-w-bush accessed 15 October 2010.

Byman, Daniel & Sarah E. Kreps, “Agents of Destruction? Applying Principal-Agent Analysis to State-Sponsored Terrorism.” International Studies Perspectives, Vol. 11, No. 1. (2010) pp. 1-18.

Byman, Daniel, Peter Chalk, Bruce Hoffman, William Rosenau, David Brannan, Trends in outside support for Insurgent Movements. Santa Monica, CA: National Security Research Division, RAND, 2001.

Byman, Daniel. Deadly Connections: States that Sponsor Terrorism. New York: Cambridge University Press, 2005.

_____. “Rogue Operators.” The National Interest (July/August 2008).

Canci, Haldun and Sevket Serkan Sen, “Turkish dilemma after Operation Desert Storm (1990-1991): an analysis of “negative consequences” Innovation—The European Journal of Social Science Research. Vol. 23, No. 3, (September 2010) pp. 279- 292.

Cann, Rebecca & Constantine Danopoulos, “The Military and Politics in a Theocratic State: Iran as Case Study,” Armed Forces and Society. 1998, No. 24

Caprioli, Mary & Peter F. Trumbore “Identifying `Rogue' States and Testing their Interstate Conflict Behavior,” European Journal of International Relations Vol. 9 (2003).

251 Carter, David B. “A Blessing or a Curse? State Support for Terrorist Groups,” International Organization, Vol. 66 No. 1 (2012): 129-151.

Ceroli, Michael A. State Sponsored Terrorism as Foreign Policy. Masters Thesis. Troy State University at Fort Bragg, North Carolina. July 1994.

Chiozza, Giacomo and H.E. Goemans, “International Conflict and the Tenure of Leaders: Is War Still Ex Post Inefficient?” American Journal of Political Science, Vol. 48, No. 3 (July 2004), pp. 604-619.

Christensen, Thomas J. Useful Adversaries: Grand Strategy, Domestic Mobilization, and Sino-American Conflict 1947-1958. Princeton, NJ: Princeton University Press, 1996.

Claiborne, William. “Iranian Embassy Staff Remains in Baghdad Despite; Iran-Iraq War is Fierce but Undeclared; DIPLOMATIC NICETY,” The Washington Post, December 1, 1980. LexisNexis Academic accessed November 5, 2012.

Cline, Ray S. and Yonah Alexander, Terrorism as State-Sponsored Covert Warfare, Fairfax: Hero Books, 1986

Collins, Stephen D. “Dissuading State Support of Terrorism: Strikes or Sanctions? (An Analysis of Dissuasion Measures Employed Against Libya)" Studies in Conflict and Terrorism, Vol. 27, No. 1 (2004) pp. 1-18.

Conybeare, J.A.C. & Todd Sandler, “State-sponsored violence as a tragedy of the commons: England’s privateering wars with France and Spain, 1625-1630,” Public Choice Issue 77. (1993) pp. 879-897.

Cooley, John K., Libyan Sandstorm. New York, NY: Holt, Rinehart and Winston, 1982.

Cooper, Tom, “Peru vs. Ecuador, Alto-Cenepa War, 1995” Central and Latin American Database September 1, 2003 available online at http://www.acig.org/artman/publish/article_164.shtml accessed December 14, 2012.

Crenshaw, Martha, "The Psychology of Terrorism: An Agenda for the 21st Century," Political Psychology Vol. 21, No. 2 (June 2000) pp. 405-420.

Crenshaw, Martha, “The Causes of Terrorism,” Comparative Politics, Vol. 13 No. 4, July 1981.

Cronin, Audrey Kurth. “Rethinking Sovereignty: American Strategy in the Age of Terrorism” Survival, Vol. 44 No. 2 (Summer 2002).

252 David, Steven R. “Explaining Third World Alignment,” World Politics Vol. 43, No.2 (January 1991), pp. 233-256.

Decker, Raymond J., S. L. C. “Report to Congressional Requesters Combating Terrorism Interagency Framework and Agency Program to Address the Overseas Threat.” U. S. G. A. Office, United State General Accounting Office, 2003.

Desch, Michael C. Civilian Control of the Military: The Changing Security Environment, Baltimore, MD: The Johns Hopkins University Press, 1999.

Donohue, Laura K. The Cost of Counterterrorism: Power, Politics, and Liberty. New York, NY: Cambridge University Press, 2008.

Downs, Anthony, An Economic Theory of Democracy. New York, NY: Harper and Collins, 1957.

Dueck, Colin. Reluctant Crusaders: Power, Culture, and Change in American Grand Strategy. Princeton, NJ: Princeton University Press, 2006.

El-Khawas, Mohamed A. Qaddafi: His Ideology in Theory and Practice. Brattleboro, VT: Maple Leaf Press, 1986: 103.

Enders, Walter & Todd Sandler, “Patterns of transnational terrorism, 1970-1999: alternative time series estimates,” International Studies Quarterly, 2002.

Enloe, Cynthia H., “The Issue Saliency of the Military-Ethnic Connection: Some thoughts on Malaysia,” Comparative Politics, Vol. 10 No. 2 (Jan 1978) pp. 267- 285.

“Evolution of Malaysia’s Foreign Policy” Ministry of Foreign Affairs, Malaysia available online at http://www.kln.gov.my/web/guest/evolution accessed December 14, 2012.

Fearon, James D. “Signaling versus the Balance of Power and Interests: An Empirical Test of a Crisis Bargaining Model,” Journal of Conflict Resolution, Vol. 38. No. 2 (June 1994) pp. 236-269.

Fernandez, Erwin S., “Philippine-Malaysia Dispute: A Bibliographic Survey” Asia- Pacific Social Science Review Volume 7 Number 1 December 2007. Pp. 53-64. Finnemore, Martha, National Interests in International Society. Ithaca, NY: Cornell University Press, 1996.

First, Ruth, The barrel of a gun: political power in Africa and the coup d’état. Part V: The Soldiers Invade: Coup Casebooks. Republished by the Ruth First Papers Project London, UK: Institute of Commonwealth Studies, 2012. Available online

253 at www.ruthfirstpapers.org.uk accessed March 8, 2013.

Fischer, Fritz. War of Illusions: German Policies from 1911-1914. New York, NY: Alfred K. Knopf, 1975.

Focus on Libya: February 1984-June 1989 A Resource Compendium. Washington, D.C.: PEMCON Ltd., 1989.

Fontaine, Roger W. "Terrorism: The Cuban Connection" New York, NY: Crane Russak & Company, 1988.

Frankel, Glenn. “Lines in the Sand,” The Gulf War Reader: History, Documents, Opinion, Micah L. Sifry and Christopher Cerf, eds. New York, NY: Times Books, 1991.

Gal-Or, Noemi, “State Sponsored Terrorism: A Mode of Diplomacy” Conflict Quarterly. (Summer 1993).

Gaucher, R. Les Terroristes. Paris, 1965.

Ghosn, Faten, Glenn Palmer, and Stuart Bremer, "The MID3 Data Set, 1993–2001: Procedures, Coding Rules, and Description," Conflict Management and Peace Science Vol. 21 (2004) pp.133-154.

Gilpin, Robert. War and Change in World Politics. New York, NY: Cambridge University Press, 1981.

Ginat, Rami and Uri Bar-Noi, "Tacit Support for Terrorism: The Rapprochement between the USSR and Palestinian Guerrilla Organizations following the 1967 War," The Journal of Strategic Studies, Vol. 30 No. 2. (April 2007) pp. 275.

Golingai, Philip “Philippine peace deal must take hold,” Asia News Network October 15, 2012. Available online at http://www.asianewsnet.net/home/news.php?sec=3&id=37637 accessed December 12, 2012.

Goulka, Jeremiah, Lydia Hansell, Elizabeth Wilkde, & Judith Larson, “The Mujahedin-e Khalq in Iraq. A Policy Conundrum” Santa Monica, CA: National Defense Research Institute, Rand Corporation, 2009.

Harris, Lillian Craig, “Libya: Qadhafi’s Revolution and the Modern State,” Boulder, CO: Westview Press: 1986.

Herspring, Dale R., The Pentagon and the Presidency: Civil-Military Relations from FDR to George W. Bush. Lawrence, Kansas: University of Kansas Press, 2005.

254 Hoffman, Bruce, Inside Terrorism. New York: Columbia University Press, 2006.

Hosmer, David W. Jr. & Stanley Lemeshow, Applied Logistic Regression, Hoboken NJ: Wiley and Sons. Inc., 2000.

“How to be a Basque on both sides of the border; by the Paris correspondent; The Economist; November 24, 1979. LexisNexis Academic (April 15, 2011).

Huntington, Samuel P., The Soldier and the State; the theory and politics of civil- military relations, New York, NY: Vintage Books, 1957.

Islamic Republic of Iran Plan and Budget Organization. Summary Report an Estimate of the Economic Damages of the Imposed War of Iraq against Iran. Tehran, Iran: Socio-economic Documents and Publications Center Plan and Budget Organization, 1983.

Jenkins, Brian. “International Terrorism: A New Kind of Warfare” RAND document P- 5261. Santa Monica, CA: RAND Corporation, June 1974.

Jinks, Derek, “State Responsibility for the Acts of Private Armed Groups,” Chicago Journal of International Law, Vol. 4 (2003) pp. 1-15.

Kegley, Charles W. Jr. T. Vance Sturgeon & Eugene R. Wittkopf. “Structural Terrorism: The Systemic Sources of State-Sponsored Terrorism” in Terrible Beyond Endurance? The Foreign Policy of State Terrorism, Michael Stohl and George A. Lopez eds. New York: Greenwood Press, 1988.

Kegley, Charles W. Jr., International Terrorism: Characteristics, Causes, and Controls. New York, NY: St. Martin’s Press, Inc., 1990.

Khalidi, Walid. “Iraq vs. Kuwait: Claims and Counterclaims,” The Gulf War Reader: History, Documents, Opinions. Micah L. Sifry and Christopher Cerf, eds. New York, NY: Times Books, 1991.

Krause-Jackson, Flavia and John Walcott, “U.S. May Buy Looted Libyan Missiles Sold in Mali’s Black Market” Aug 25, 2011 available from http://www.bloomberg.com/news/2011-08-25/u-s-may-snap-up-qaddafi-s- missiles-in-mali-black-market-to-soak-up-supply.html accessed August 31, 2011.

Lacquer, Walter & Yonah Alexander, "Libya Under Qadhafi: a pattern of aggression" in The Terrorism Reader. Walter Lacquer & Yonah Alexander Eds. New York, NY: Meridian Books: 1987.

255 LaFree, Gary, and Laura Dugan. Global Terrorism Database II, 1998-2004 [Computer file]. ICPSR22600-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2008-10-23. doi:10.3886/ICPSR22600

Lassinantti, Gunner. “Small States and Alliances—A Swedish Perspective.” In Small States and Alliances, Erich Reiter and Heinz Gartner, eds. Springer, 2001

Lemke, Douglas and William Reed "The Relevance of Politically Relevant Dyads," Journal of Conflict Resolution Vol. 45, No. 1 (Feb. 2001) pp. 126-144.

Lesser, Ian O., “Countering the New Terrorism: Implications for Strategy,” in Countering the New Terrorism, Ian O. Lesser, Bruce Hoffman, John Arquilla, David Ronfeldt, Michele Zanini eds. Santa Monica, CA: Project Air Force, RAND corporation, 1999.

Levitt, Matthew. Targeting terror: U.S. policy toward Middle Eastern state sponsors and terrorist organizations, post-September 11. Washington, DC: Washington Institute for Near East Policy, 2002.

Livingstone, Neil C. & Terrell E. Arnold, eds. Fighting Back: Winning the War against Terrorism. Lexington, MA: Lexington Books, 1986.

Long, J. Scott & Jeremy Freese, Regression Models for Categorical Dependent Variables Using Stata, Second Edition, College Station, TX: Stata Press, 2006.

Maddala, G.S., “Limited Dependent Variable Models Using Panel Data,” The Journal of Human Resources. Vol. 22, No. 3 (Summer, 1987) pp. 307-338.

Malek, Martin, “Russia, Iran, and the Conflict in Chechnya,” Caucasian Review of International Affairs, Vol. 1 (2) Winter 2008, pp. 25-34.

Maoz, Zeev. Domestic Sources of Global Change, Ann Arbor, MI: University of Michigan Press, 1996.

____. National Choices and International Processes. Cambridge, UK: Cambridge University Press, 1990.

Maoz, Zeev and Bruce Russett, "Normative and Structural Causes of Democratic Peace: 1946-1986," American Political Science Review, Vol. 87, No. 3 (1993) pp. 624- 638.

Marshall, Monty G. and Donna Ramsey Marshall, "Coup d’état Events," Center for Systemic Peace, 2010. Available online at the Integrated Network for Societal Conflict Research Data Page http://www.systemicpeace.org/inscr/inscr.htm accessed December 16, 2012.

256 Mauceri, Philip, “State Reform, Coalitions, and the Neoliberal Autoglope in Peru,” Latin American Research Review, Vol. 30, No. 1 (1995), pp. 7-37.

McGillivray, Fiona & Alastair Smith. Punishing the Prince: a theory of interstate relations, political institutions, and leader change. Princeton, NJ: Princeton University Press, 2008.

Mckinney, Jack. “Kidnap a Soviet? Well…Shiite Moslems tried it in Lebanon—Once.” The Philadelphia Inquirer, February 26, 1988. Available online at http://articles.philly.com/1988-02-26/news/26242928_1_islamic-liberation- organization-soviets-hezbollah accessed March 6, 2013.

Mearshimer, John J. The Tragedy of Great Power Politics, New York, NY: Norton, 2001.

Melman, Yossi, "Iraq's Ties to Terror: The Threat Isn't Easy to Read" New York Times; February 9, 2003;

Merari, Ariel. “Terrorism as a Strategy of Insurgency,” Terrorism and Political Violence, Vol. 5, No. 4 (Winter 1993) pp. 213-251.

Mesbahi, Mohiaddin. “The USSR and the Iran-Iraq War” The Iran-Iraq War: The Politics of Aggression. Farhang Rajaee, ed. Gainesville, FL: University Press of Florida, 1993.

Metz, Helen Chapin, ed. Iraq: A Country Study. Washington: GPO for the Library of Congress, 1988. Available online at http://countrystudies.us/iraq/99.htm, accessed March 5, 2013.

Mickolus, Edward. “What Constitutes State Support of Terrorists?” Terrorism and Political Violence, 1(3) July 1989.

Mickolus, Edward, Todd Sandler & Jean M. Murdock, International Terrorism in the 1980s: A Chronology of Events. Ames, Iowa: Iowa State University Press, 1989.

Miller, Erin and Kathleen Smarick, 2012, "Profiles of Perpetrators of Terrorism in the United States (PPT-US)", http://hdl.handle.net/1902.1/17702 UNF:5:rMan9QGKaQXo4bRl8luCzQ== National Consortium for the Study of Terrorism and Responses to Terrorism [Distributor] V3 [Version]

Milne, R.S. and Diane K. Mauzy, Politics and Government in Malaysia. Singapore: Federal Publications (S) Pte Ltd, 1978.

Moghadam, Assaf. The Globalization of Martyrdom: Al Qaeda, Salafi Jihad, and the Diffusion of Suicide Attacks. Baltimore: Johns Hopkins University Press, 2008.

257 Mohindra, Satish. Terrorist Games Nations Play, New Delhi, India: Lance Publishers; Hartford, WI, USA: Spantech & Lancer, 1993.

Monti-Belkaoui, Janice and Ahmed Riahi-Belkaoui, Qaddafi: The Man and His Policies, Brookfield, VT: Ashgate Publishing Company, 1996.

“Montesinos: The end of the road” BBC News, June 24, 2001 available online at http://news.bbc.co.uk/2/hi/americas/1405548.stm accessed December 14, 2012

Moore, John Norton ed. Deception and Deterrence in “Wars of National Liberation State-sponsored Terrorism and other forms of secret warfare.” Durham, NC: Carolina Academic Press, 1997.

“Moro National Liberation Front” Federation of Atomic Scientists, Intelligence Resource Program. Available online at http://www.fas.org/irp/world/para/mnlf.htm accessed October 20, 2012.

Morrow, James D. “Alliances and Asymmetry: An Alternative to the Capability Aggregation Model of Alliances” American Journal of Political Science, Vol. 35, No. 4 (November 1991).

Murphy, J. “Role of international law in the prevention of terrorist kidnapping of diplomatic personnel.” International Terrorism and Political Crimes. Springfield, IL: Thomas, 1976.

Nacos, BL. Terrorism and counterterrorism: understanding threats and responses in the post-9/11 world. Boston, MA: Longman. 2010.

National Consortium for the Study of Terrorism and Responses to Terrorism (START). (2012). Global Terrorism Database [Data file]. Retrieved from http://www.start.umd.edu/gtd accessed July 9, 2012.

Neuberger, Ralph Benyamin, Involvement, Invasion and Withdrawal: Qadhdhafi’s Libya and Chad, 1969-1981, Tel-Aviv, Israel: Shiloah Center for Middle Eastern and African Studies: 1982, p. 61; 29;

Nincic, Miroslav. Renegade Regimes: Confronting Deviant Behavior in World Politics. New York, NY: Colombia University Press, 2005.

Nobel, Lela Garner, “The Moro National Liberation Front in the Philippines,” Pacific Affairs, Vol. 49, No. 3 (Autumn, 1976) pp. 405-424.

Norden, Deborah L. “Democracy and Military Control in Venezuela: From Subordination to Insurrection” Latin American Research Review, Vol. 33, No. 2 (1998), pp. 143- 1665.

258 Olson, Mancur. Rise and Decline of Nations. New Haven, CT: Yale University Press, 1982.

Origin Unknown, “Peru: Military Unease Growing,” December 23, 1992. Classification Unknown. Available online at http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB37/ accessed November 2, 1992.

Osgood, Robert E. and Robert W. Tucker. Force, Order, and Justice. Baltimore: Johns Hopkins Press, 1967.

Overgaard, Per Baltzer, “The Scale of Terrorist Attacks as a Signal of Resources,” The Journal of Conflict Resolution, Vol. 38, No. 3 (Sept. 1994) pp. 452-478.

Palmer, David Scott, “Peru-Ecuador Border Conflict: Missed Opportunities, Misplaced Nationalism, and Multilateral Peacekeeping,” Journal of Interamerican Studies and World Affairs. Vol. 39, No. 3 (Autumn 1997): 109-148.

Pape, Robert. “The Strategic Logic of Suicide Terrorism.” American Political Science Review, Vol. 97 No. 3 (2003).

Patrick, S. ”Weak States ad Global Threats: Fact or Fiction?” The Washington Quarterly Vol. 29 No. 2 (2006): 27-53.

PERUPAZ, “Atentados,” (Lima: Instituto Consitucion y Sociedad), March 1993 p. 9; cited in Alberto Arciniega Huby. “Documentation: Civil-Military Relations and a Democratic Peru,” Orbis, Vol. 38, No.1 (Winter 1994) pp. 109-117.

Piazza, James A. “Incubators of Terror: Do Failed and Failing States Promote Transnational Terrorism?” International Studies Quarterly, Vol. 52 No. 3. (2008).

Pillar, Paul. Terrorism and US Foreign Policy. Brookings Institution Press. 2004.

Polity IV Project: Dataset Users’ Manual: 15. Available from http://www.systemicpeace.org/inscr/inscr.htm accessed 5 November 2009

Prunckun, Henry W. “Military Deterrence of International Terrorism: An Evaluation of Operation El Dorado Canyon,” Studies in Conflict & Terrorism, Vol. 20, No. 3 (July-Sept. 1997) pp. 267-281.

Prussian, Nathan J., Countering Iranian Sponsored Terrorism, Masters Thesis, Joint Forces Staff College, Joint Advanced Warfighting School, 10 June 2010.

Putnam, Robert D. “Diplomacy and Domestic Politics: The Logic of Two-Level Games.” International Organization Vol. 42, (1988): 427-60.

259

Quillen, Chris. “A Historical Analysis of Mass Casualty Bombers” Studies in Conflict and Terrorism. Vol. 25, No. 5 (September 2002) pp. 279-292.

Rajaee, Farhang. “The Iran-Iraq War: The Politics of Aggression” University Press of Florida, 1993.

Ramazani, R.K., Revolutionary Iran: Challenge and Response in the Middle East, Baltimore, MN: Johns Hopkins University Press 1985: 44

Randal, Johnathan C., “Iran Stages Anti-US Rally; Firing squads execute 13 more” The Washington Post. March 13, 1979. Available from LexisNexis Academic, accessed March 4, 2013.

Rasler, Karen A. and William R. Thompson, "Replication data for: Contested Territory, Strategic Rivalries and Conflict Escalation,” Published in International Studies Quarterly Vol. 50 No.1 (2006) pp. 145–168 available online at http://de.scientificcommons.org/48542826.

Reagan, Ronald. National Security Decision Directive 141, May 25, 1984. Declassified. Available online at http://www.fas.org/irp/offdocs/nsdd/nsdd-141.pdf accessed March 6, 2013.

Report of the Panel of Experts on Violations of Security Council Sanctions against UNITA, UNSC S/2000/203 May 7, 1999 http://www.un.org/News/dh/latest/angolareport_eng.htm accessed 12 August 2011.

Rice, C. “Rethinking the National Interest” Foreign Affairs Vol. 87 No. 4 (2008): 2-26.

Richards, Paul. “New War: An Ethnographic Approach” No Peace No War: An Anthropology of Contemporary Armed Conflicts. Paul Richards ed. Athens, Ohio: Ohio University Press: 2005.

Ricks, Thomas E. “Containing Iraq: A Forgotten War,” The Washington Post, October 25, 2000. Available online at http://www.globalpolicy.org/component/content/article/168/34646.html accessed March 15, 2013.

Ripsman, Norrin M. and Jack S. Levy. “Wishful Thinking or Buying Time? The Logic of British Appeasement in the 1930s.” International Security Vol. 33, No. 2 (Fall 2008): 148-181.

Ronen, Yehudit, Qaddafi’s Libya in World Politics Lynne Reinner Publishers, Inc. 2008: 151

260 Rotberg, Robert, “The New Nature of Nation-State Failure,” The Washington Quarterly, Vol. 25 No. 3. (2002): 85-96.

Rubin, Barry. “US Foreign Policy and Rogue States,” Middle East Review of International Affairs Journal, Vol. 3, No. 3 (September 1999).

Salahyan, Idean “Transnational Rebels: Neighboring States as Sanctuary for Rebel Groups” World Politics Vol. 59 (January 2007): 217-242.

Salahyan, Idean, Kristian Gleditsch, and David Cunningham. Non State Actor Dataset, Version 2.7 2009 available online through http://privatewww.essex.ac.uk/~ksg/eacd.html accessed December 16, 2012.

Samad, Paridah Abd and Darusalam Abu Bakar, "Malaysia-Philippines Relations: The Issue of Sabah" Asian Survey Vol. 32. No. 6 (Jun 1992) pp. 556.

San Akca, Belgin. “Non-State Violence, Supporters, and Targets: State Support of Non- State Armed Groups as a Selection Process” Dissertation, University of California-Davis, 2009.

Sandler, Todd and Walter Enders, "An Economic Perspective on Transnational Terrorism," European Journal of Political Economy, Vol. 20 (2004) pp. 301-316.

Schelling, Thomas. Arms and Influence. New Haven, CT: Yale University Press. 1966.

Schmid, Alex P. and Albert J. Jongman. Political Terrorism: A New Guide to Actors, Authors, Concepts, Data Bases, Theories, & Literature. New Brunswick: Transaction Publishers, 1988.

Schultz, Kenneth A. Democracy and Coercive Diplomacy. Cambridge, UK: Cambridge University Press, 2001.

Schultz, Richard “Conceptualizing Political Terrorism: A Typology,” in International Terrorism: Current Research and Directions. Alan D. Bucklye and Daniel D. Olson, eds., Wayne, NJ: Avery Publishing Group, Inc., 1980.

Seale, Patrick. “Abu Nidal A Gun for Hire: The Secret Life of the World’s Most Notorious Arab Terrorist. New York, NY: Random House Publishing, 1992.

Selth, Andrew. Against every Human Law: the Terrorist Threat to Diplomacy. Rushcutters Bay, N.S.W., Australia: Australian National University Press, 1988.

Shaw, William. Tun Razak, His Life and Times. Wisma Damansara, Jalan Sematan, Kuala Lumpur: Longman Malaysia SDN Berhard, 1976.

261 Shay, Shaul, The Axis of Evil: Iran, Hizballah, and Palestinian Terror. New Brunswick, NJ: Transaction Publishers, 2005.

Sick, Gary, “October Surprise: America’s Hostages in Iran and the Election of Ronald Reagan” New York, NY: Random House Inc.: 1991: 74-5.

_____. “Trial by Error: Reflections on the Iran-Iraq War,” The Middle East Journal. Vol. 43, No. 2 (Spring 1989), pp. 230-245.

Silke, Andrew. Research on Terrorism: Trends, Achievements, and Failures. New York, NY: Taylor and Francis, 2004.

Siqueira, Kevin and Todd Sandler, “Terrorists versus the Government: Strategic interaction, Support, and Sponsorship. Journal of Conflict Resolution, Vol. 50 (2006) pp. 878-897.

Slann, Martin W. and Bernard Schechterman. Multidimensional Terrorism. Boulder, CO: L. Rienner, 1987.

Smith, M.L.R. “Strategy in an age of ‘Low-Intensity’ Warfare,” Rethinking the Nature of War, Isabelle Duyvesteyn & Jan Angstrom, eds. New York, NY: Routledge, 2004.2005.

Snyder, Jack. Myths of Empire: Domestic Politics and International Ambition. Ithaca, NY: Cornell University Press, 1991.

St John, Ronald Bruce, Libya: From Colony to Independence. Oxford, England: Oneworld Publications, 2008.

_____. Qaddafi’s World Design: Libyan Foreign Policy 1969-1987. Atlantic Highlands, NJ: Saqi Books, 1987: 37.

Sterling, Claire. "The Terror Network: The Secret War of International Terrorism" New York, NY: Holt, Rinehart and Winston, 1981.

Stockholm International Peace Research Institute (SIPRI), Arms trade database Available from http://www.sipri.org/research/armaments/transfers accessed 20 August 2011.

Stohl, Michael & George A Lopez. The State as Terrorist: The Dynamics of Governmental Violence and Repression. Greenwood Pub Group, 1984.

Stohl, Michael & George A. Lopez eds. Terrible Beyond Endurance? The Foreign Policy of State Terrorism. New York: Greenwood Press, 1988.

262 Strong, Simon, Shining Path: The Worlds Deadliest Revolutionary Force. London, UK: HarperCollins Publishers, 1992.

Testimony of Howard Teicher. United States v. Cardoen No: 93-241 (Southern District Court of Florida, Jan. 31, 1995). http://www.informationclearinghouse.info/article1413.htm accessed March 5, 2013.

“The 1995 Peruvian-Ecuadorian Border Conflict,” NativeWeb February 1995, available online at http://abyayala.nativeweb.org/ecuador/border/border1.html accessed December 14, 2012.

The Military Balance, London: International Institute for Strategic Studies. Volumes 1969/70; 1970/71; 1971/72; 1972/73; 1973/74; 1974/75; 1975/76; 1976/77; 1977/78; 1978/79; 1979/80; 1980/81; 1981/82; 1982/83; 1983/84; 1984/85; 1985/86; 1986/87; 1987/88; 1988/89; 1989/90; 1990/91; 1991/92; 1992/93; 1993/94; 1994/95; 1995/96; 1996/97; 1997/98; 1998/99; 1999/2000; 2000/01; 2001/02.

Thompson, Janice E. Mercenaries, Pirates & Sovereigns: State-building and Extraterritorial violence in Early Modern Europe. Princeton, NJ: Princeton University Press, 1994.

Thompson, Loren B. Low-intensity conflict: the Pattern of Warfare in the Modern World. Lexington, MA: Lexington Books, 1989.

Trubowitz, Peter. Politics and Strategy: Partisan Ambition & American Statecraft. Princeton, NJ: Princeton University Press, 2010.

UCDP External support dataset available at http://www.pcr.uu.se/research/ucdp/datasets/ucdp_external_support_data/ accessed 10 January 2012.

U.S. Department of State “Chapter 3: State Sponsors of Terrorism Overview” Country Reports on Terrorism, 2007. Available from http://www.state.gov/s/ct/rls/crt/2007/103711.htm accessed 26 September 2011.

_____. “State Sponsors of Terrorism” http://www.state.gov/s/ct/c14151.htm accessed 26 Sept. 2011.

_____. Patterns of Global Terrorism: 1998. Available online at http://www.state.gov/www/global/terrorism/1998Report/1998index.html accessed 26 September 2011.

263 _____. “Chapter 3: State Sponsors of Terrorism,” Country Reports on Terrorism, 2011 (July 31, 2012) available online at http://www.state.gov/j/ct/rls/crt/2011/195547.htm accessed November 5, 2012.

Van Evera, Stephen. Causes of War: Power and the Roots of Conflict. Ithaca, NY: Cornell University Press, 1999.

Vinci, Anthony. Armed Groups and the Balance of Power: The International Relations of Terrorists, Warlords, and Insurgents. New York, NY: Routledge, 2009.

Volgy, Thomas J., Lawrence E. Imwalle, and Jeff J. Corntassel.“Structural Determinants of International Terrorism: The Effects of Hegemony and Polarity on Terrorist Activity.” International Interactions Vol. 23 (1997): 207-31.

Walt, Stephen M. Revolution and War. Ithaca, NY: Cornell University, 1996.

_____. The Origins of Alliances. Ithaca, NY: Cornell University Press, 1987.

Waltz, Kenneth N., Theory of International Politics, Boston, MA: McGraw Hill, 1979.

Weber, Max. Politics as a Vocation, Philadelphia, PA: Fortress Press, 1965.

Wehrey, Frederic, Jerrold D. Green, Brian Nichiporuk, Alireza Nader, Lydia Hansell, Rasool Nafisi, & S.R. Bohandy. The Rise of the Pasadran: Assessing the Domestic Roles of Iran’s Islamic Revolutionary Guards Corps. Santa Monica, CA: RAND National Defense Research Institute, 2009.

Weinberg, Leonard & Ami Pedahzur. Political Parties and Terrorist Groups. New York, NY: Routledge, 2003.

Weinberg, Leonard. Democratic Responses to Terrorism, New York, NY: Routledge, 2008.

Wilkinson, Paul. Terrorism and the Liberal State, Hoboken, NJ: Halsted Press, 1977.

______. Political Terrorism. London: Macmillan, 1974.

______. Terrorism Versus Democracy: The Liberal State Response, second edition revised and updated. New York, NY: Routledge, 2006.

Yaffee, Robert, “A Primer for Panel Data” 2003 available online at http://www.nyu.edu/its/pubs/connect/fall03/yaffee_primer.html accessed Sept. 6, 2012.

Yallop, David, Tracking the Jackal: The Search for Carlos, the World's Most Wanted Man. New York, NY: Random House, 1993.

264

Zack-Williams, A. & Stephen Riley. “Sierra Leone: The Coup and its Consequences” Review of African Political Economy. No. 56 (March 1993), pp. 91-98.

Zakaria, Fareed. From Wealth to Power: The Unusual Origins of America’s World Role. Princeton, NJ: Princeton University Press, 1997.

265